Project: Identify Customer Segments¶

In this project, you will apply unsupervised learning techniques to identify segments of the population that form the core customer base for a mail-order sales company in Germany. These segments can then be used to direct marketing campaigns towards audiences that will have the highest expected rate of returns. The data that you will use has been provided by our partners at Bertelsmann Arvato Analytics, and represents a real-life data science task.

This notebook will help you complete this task by providing a framework within which you will perform your analysis steps. In each step of the project, you will see some text describing the subtask that you will perform, followed by one or more code cells for you to complete your work. Feel free to add additional code and markdown cells as you go along so that you can explore everything in precise chunks. The code cells provided in the base template will outline only the major tasks, and will usually not be enough to cover all of the minor tasks that comprise it.

It should be noted that while there will be precise guidelines on how you should handle certain tasks in the project, there will also be places where an exact specification is not provided. There will be times in the project where you will need to make and justify your own decisions on how to treat the data. These are places where there may not be only one way to handle the data. In real-life tasks, there may be many valid ways to approach an analysis task. One of the most important things you can do is clearly document your approach so that other scientists can understand the decisions you've made.

At the end of most sections, there will be a Markdown cell labeled Discussion. In these cells, you will report your findings for the completed section, as well as document the decisions that you made in your approach to each subtask. Your project will be evaluated not just on the code used to complete the tasks outlined, but also your communication about your observations and conclusions at each stage.

In [ ]:
In [2601]:
# import libraries here; add more as necessary
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
import re
from scipy import stats


from sklearn.preprocessing import OneHotEncoder, StandardScaler
from sklearn.impute import SimpleImputer
from sklearn.decomposition import PCA
from sklearn.cluster import KMeans
from sklearn.metrics import silhouette_score


from collections import Counter


from yellowbrick.cluster import KElbowVisualizer


pd.options.display.max_rows = None
pd.options.display.max_columns = None
pd.set_option("future.no_silent_downcasting", True)
In [1912]:
!pip install yellowbrick

Requirement already satisfied: yellowbrick in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (1.5)
Requirement already satisfied: matplotlib!=3.0.0,>=2.0.2 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from yellowbrick) (3.8.4)
Requirement already satisfied: scipy>=1.0.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from yellowbrick) (1.13.1)
Requirement already satisfied: scikit-learn>=1.0.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from yellowbrick) (1.5.1)
Requirement already satisfied: numpy>=1.16.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from yellowbrick) (1.26.4)
Requirement already satisfied: cycler>=0.10.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from yellowbrick) (0.11.0)
Requirement already satisfied: contourpy>=1.0.1 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (1.2.0)
Requirement already satisfied: fonttools>=4.22.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (4.51.0)
Requirement already satisfied: kiwisolver>=1.3.1 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (1.4.4)
Requirement already satisfied: packaging>=20.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (24.1)
Requirement already satisfied: pillow>=8 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (10.4.0)
Requirement already satisfied: pyparsing>=2.3.1 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (3.0.9)
Requirement already satisfied: python-dateutil>=2.7 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from matplotlib!=3.0.0,>=2.0.2->yellowbrick) (2.9.0.post0)
Requirement already satisfied: joblib>=1.2.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from scikit-learn>=1.0.0->yellowbrick) (1.4.2)
Requirement already satisfied: threadpoolctl>=3.1.0 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from scikit-learn>=1.0.0->yellowbrick) (3.5.0)
Requirement already satisfied: six>=1.5 in /Users/chrismo/opt/anaconda3/lib/python3.11/site-packages (from python-dateutil>=2.7->matplotlib!=3.0.0,>=2.0.2->yellowbrick) (1.16.0)
In [ ]:

Step 0: Load the Data¶

There are four files associated with this project (not including this one):

  • Udacity_AZDIAS_Subset.csv: Demographics data for the general population of Germany; 891211 persons (rows) x 85 features (columns).
  • Udacity_CUSTOMERS_Subset.csv: Demographics data for customers of a mail-order company; 191652 persons (rows) x 85 features (columns).
  • Data_Dictionary.md: Detailed information file about the features in the provided datasets.
  • AZDIAS_Feature_Summary.csv: Summary of feature attributes for demographics data; 85 features (rows) x 4 columns

Each row of the demographics files represents a single person, but also includes information outside of individuals, including information about their household, building, and neighborhood. You will use this information to cluster the general population into groups with similar demographic properties. Then, you will see how the people in the customers dataset fit into those created clusters. The hope here is that certain clusters are over-represented in the customers data, as compared to the general population; those over-represented clusters will be assumed to be part of the core userbase. This information can then be used for further applications, such as targeting for a marketing campaign.

To start off with, load in the demographics data for the general population into a pandas DataFrame, and do the same for the feature attributes summary. Note for all of the .csv data files in this project: they're semicolon (;) delimited, so you'll need an additional argument in your read_csv() call to read in the data properly. Also, considering the size of the main dataset, it may take some time for it to load completely.

Once the dataset is loaded, it's recommended that you take a little bit of time just browsing the general structure of the dataset and feature summary file. You'll be getting deep into the innards of the cleaning in the first major step of the project, so gaining some general familiarity can help you get your bearings.

In [1916]:
# Load in the general demographics data.
pop = pd.read_csv('Udacity_AZDIAS_Subset.csv', sep=';')

# Load in the feature summary file.
feat_info = pd.read_csv('AZDIAS_Feature_Summary.csv', sep=';')
In [1918]:
# Checking the structure of the data after it's loaded (i.e. print the number of
# rows and columns, print the first few rows).
In [1920]:
print(f'The Population dataset has {pop.shape[0]} rows and {pop.shape[1]} columns')

The Population dataset has 891221 rows and 85 columns
In [1922]:
print(f'The Features dataset has {feat_info.shape[0]} rows and {feat_info.shape[1]} columns')

The Features dataset has 85 rows and 4 columns
In [1924]:
pop.head()
Out[1924]:
AGER_TYP ALTERSKATEGORIE_GROB ANREDE_KZ CJT_GESAMTTYP FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER FINANZTYP GEBURTSJAHR GFK_URLAUBERTYP GREEN_AVANTGARDE HEALTH_TYP LP_LEBENSPHASE_FEIN LP_LEBENSPHASE_GROB LP_FAMILIE_FEIN LP_FAMILIE_GROB LP_STATUS_FEIN LP_STATUS_GROB NATIONALITAET_KZ PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV SHOPPER_TYP SOHO_KZ TITEL_KZ VERS_TYP ZABEOTYP ALTER_HH ANZ_PERSONEN ANZ_TITEL HH_EINKOMMEN_SCORE KK_KUNDENTYP W_KEIT_KIND_HH WOHNDAUER_2008 ANZ_HAUSHALTE_AKTIV ANZ_HH_TITEL GEBAEUDETYP KONSUMNAEHE MIN_GEBAEUDEJAHR OST_WEST_KZ WOHNLAGE CAMEO_DEUG_2015 CAMEO_DEU_2015 CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_ANTG3 KBA05_ANTG4 KBA05_BAUMAX KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP KBA13_ANZAHL_PKW PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_BAUMAX PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB
0 -1 2 1 2.0 3 4 3 5 5 3 4 0 10.0 0 -1 15.0 4.0 2.0 2.0 1.0 1.0 0 0 5.0 2 6 7 5 1 5 3 3 4 7 6 6 5 3 -1 NaN NaN -1 3 NaN NaN NaN 2.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
1 -1 1 2 5.0 1 5 2 5 4 5 1 1996 10.0 0 3 21.0 6.0 5.0 3.0 2.0 1.0 1 14 1.0 5 4 4 3 1 2 2 3 6 4 7 4 7 6 3 1.0 0.0 2 5 0.0 2.0 0.0 6.0 NaN 3.0 9.0 11.0 0.0 8.0 1.0 1992.0 W 4.0 8 8A 51 0.0 0.0 0.0 2.0 5.0 1.0 6.0 3.0 8.0 3.0 2.0 1.0 3.0 3.0 963.0 2.0 3.0 2.0 1.0 1.0 5.0 4.0 3.0 5.0 4.0
2 -1 3 2 3.0 1 4 1 2 3 5 1 1979 10.0 1 3 3.0 1.0 1.0 1.0 3.0 2.0 1 15 3.0 4 1 3 3 4 4 6 3 4 7 7 7 3 3 2 0.0 0.0 1 5 17.0 1.0 0.0 4.0 NaN 3.0 9.0 10.0 0.0 1.0 5.0 1992.0 W 2.0 4 4C 24 1.0 3.0 1.0 0.0 0.0 3.0 2.0 4.0 4.0 4.0 2.0 3.0 2.0 2.0 712.0 3.0 3.0 1.0 0.0 1.0 4.0 4.0 3.0 5.0 2.0
3 2 4 2 2.0 4 2 5 2 1 2 6 1957 1.0 0 2 0.0 0.0 0.0 0.0 9.0 4.0 1 8 2.0 5 1 2 1 4 4 7 4 3 4 4 5 4 4 1 0.0 0.0 1 3 13.0 0.0 0.0 1.0 NaN NaN 9.0 1.0 0.0 1.0 4.0 1997.0 W 7.0 2 2A 12 4.0 1.0 0.0 0.0 1.0 4.0 4.0 2.0 6.0 4.0 0.0 4.0 1.0 0.0 596.0 2.0 2.0 2.0 0.0 1.0 3.0 4.0 2.0 3.0 3.0
4 -1 3 1 5.0 4 3 4 1 3 2 5 1963 5.0 0 3 32.0 10.0 10.0 5.0 3.0 2.0 1 8 5.0 6 4 4 2 7 4 4 6 2 3 2 2 4 2 2 0.0 0.0 2 4 20.0 4.0 0.0 5.0 1.0 2.0 9.0 3.0 0.0 1.0 4.0 1992.0 W 3.0 6 6B 43 1.0 4.0 1.0 0.0 0.0 3.0 2.0 5.0 1.0 5.0 3.0 3.0 5.0 5.0 435.0 2.0 4.0 2.0 1.0 2.0 3.0 3.0 4.0 6.0 5.0
In [1926]:
feat_info.head()
Out[1926]:
attribute information_level type missing_or_unknown
0 AGER_TYP person categorical [-1,0]
1 ALTERSKATEGORIE_GROB person ordinal [-1,0,9]
2 ANREDE_KZ person categorical [-1,0]
3 CJT_GESAMTTYP person categorical [0]
4 FINANZ_MINIMALIST person ordinal [-1]
In [1928]:
def check_value_counts(pop:pd.DataFrame=pop)-> None:
    """
    INPUT: pop dataframe
    OUTPUT: examines values for each column. This helps us check if our cleaning pipeline was successful
    """
    for col in pop.columns:
        print(pop[col].value_counts())
In [1930]:
#check_value_counts(pop)

Tip: Add additional cells to keep everything in reasonably-sized chunks! Keyboard shortcut esc --> a (press escape to enter command mode, then press the 'A' key) adds a new cell before the active cell, and esc --> b adds a new cell after the active cell. If you need to convert an active cell to a markdown cell, use esc --> m and to convert to a code cell, use esc --> y.

Step 1: Preprocessing¶

Step 1.1: Assess Missing Data¶

The feature summary file contains a summary of properties for each demographics data column. You will use this file to help you make cleaning decisions during this stage of the project. First of all, you should assess the demographics data in terms of missing data. Pay attention to the following points as you perform your analysis, and take notes on what you observe. Make sure that you fill in the Discussion cell with your findings and decisions at the end of each step that has one!

Step 1.1.1: Convert Missing Value Codes to NaNs¶

The fourth column of the feature attributes summary (loaded in above as feat_info) documents the codes from the data dictionary that indicate missing or unknown data. While the file encodes this as a list (e.g. [-1,0]), this will get read in as a string object. You'll need to do a little bit of parsing to make use of it to identify and clean the data. Convert data that matches a 'missing' or 'unknown' value code into a numpy NaN value. You might want to see how much data takes on a 'missing' or 'unknown' code, and how much data is naturally missing, as a point of interest.

As one more reminder, you are encouraged to add additional cells to break up your analysis into manageable chunks.

In [1933]:
feat_info['missing_or_unknown'].value_counts()
Out[1933]:
missing_or_unknown
[-1]        26
[-1,9]      17
[-1,0]      16
[0]         12
[]          10
[-1,0,9]     1
[-1,X]       1
[XX]         1
[-1,XX]      1
Name: count, dtype: int64

The 'missing_or_unknown' explains what values of our Population (pop) columns were incorrectly encoded, hence we should make them np.nans

In [1936]:
print(f'The total number of missing or unknown values in our Populations data is {pop.isna().sum().sum()}')

The total number of missing or unknown values in our Populations data is 4896838
In [1938]:
# Identifying and converting missing_or_unknown data.

for x in range(len(feat_info)):
    # using regex to extract the digits
    missing_data = re.sub('[\[|\]]', '',feat_info.iloc[x]['missing_or_unknown']).split(',')
    # return the digit values for integer strings only
    if missing_data != ['']:
        missing_data = [np.int64(data) if (data!='X' and data!='XX') else data for data in missing_data]
        pop = pop.replace({feat_info.iloc[x]['attribute']: missing_data}, np.nan)
In [1939]:
print(f'Now, the value of missing or unknown values in our Populations data, now is {pop.isna().sum().sum()}')

Now, the value of missing or unknown values in our Populations data, now is 8373929
In [1940]:
#check_value_counts(pop)

Reconfirming the training pipeline worked effectively, all looks good !

Step 1.1.2: Assess Missing Data in Each Column¶

How much missing data is present in each column? There are a few columns that are outliers in terms of the proportion of values that are missing. You will want to use matplotlib's hist() function to visualize the distribution of missing value counts to find these columns. Identify and document these columns. While some of these columns might have justifications for keeping or re-encoding the data, for this project you should just remove them from the dataframe. (Feel free to make remarks about these outlier columns in the discussion, however!)

For the remaining features, are there any patterns in which columns have, or share, missing data?

In [1946]:
missing = ((pop.isna().sum()/pop.shape[0])*100).sort_values(ascending=False)
missing
Out[1946]:
TITEL_KZ                 99.757636
AGER_TYP                 76.955435
KK_KUNDENTYP             65.596749
KBA05_BAUMAX             53.468668
GEBURTSJAHR              44.020282
ALTER_HH                 34.813699
KKK                      17.735668
REGIOTYP                 17.735668
W_KEIT_KIND_HH           16.605084
KBA05_ANTG1              14.959701
KBA05_ANTG2              14.959701
KBA05_ANTG3              14.959701
KBA05_ANTG4              14.959701
KBA05_GBZ                14.959701
MOBI_REGIO               14.959701
PLZ8_ANTG3               13.073637
PLZ8_ANTG2               13.073637
PLZ8_GBZ                 13.073637
PLZ8_HHZ                 13.073637
PLZ8_ANTG1               13.073637
PLZ8_BAUMAX              13.073637
PLZ8_ANTG4               13.073637
VERS_TYP                 12.476816
HEALTH_TYP               12.476816
SHOPPER_TYP              12.476816
NATIONALITAET_KZ         12.153551
PRAEGENDE_JUGENDJAHRE    12.136608
KBA13_ANZAHL_PKW         11.871354
ANZ_HAUSHALTE_AKTIV      11.176913
CAMEO_INTL_2015          11.147852
CAMEO_DEU_2015           11.147852
CAMEO_DEUG_2015          11.147852
LP_LEBENSPHASE_FEIN      10.954859
ARBEIT                   10.926022
RELAT_AB                 10.926022
ORTSGR_KLS9              10.914689
ANZ_HH_TITEL             10.884842
LP_LEBENSPHASE_GROB      10.611509
INNENSTADT               10.518154
EWDICHTE                 10.518154
BALLRAUM                 10.518154
GEBAEUDETYP_RASTER       10.452514
WOHNLAGE                 10.451729
OST_WEST_KZ              10.451729
MIN_GEBAEUDEJAHR         10.451729
GEBAEUDETYP              10.451729
LP_FAMILIE_FEIN           8.728699
LP_FAMILIE_GROB           8.728699
KONSUMNAEHE               8.299737
WOHNDAUER_2008            8.247000
SOHO_KZ                   8.247000
ANZ_TITEL                 8.247000
ANZ_PERSONEN              8.247000
HH_EINKOMMEN_SCORE        2.058749
GFK_URLAUBERTYP           0.544646
CJT_GESAMTTYP             0.544646
LP_STATUS_FEIN            0.544646
LP_STATUS_GROB            0.544646
RETOURTYP_BK_S            0.544646
ONLINE_AFFINITAET         0.544646
ALTERSKATEGORIE_GROB      0.323264
FINANZ_UNAUFFAELLIGER     0.000000
FINANZTYP                 0.000000
FINANZ_HAUSBAUER          0.000000
GREEN_AVANTGARDE          0.000000
FINANZ_SPARER             0.000000
FINANZ_MINIMALIST         0.000000
FINANZ_VORSORGER          0.000000
FINANZ_ANLEGER            0.000000
ANREDE_KZ                 0.000000
SEMIO_KAEM                0.000000
SEMIO_SOZ                 0.000000
SEMIO_PFLICHT             0.000000
SEMIO_FAM                 0.000000
SEMIO_REL                 0.000000
SEMIO_MAT                 0.000000
SEMIO_VERT                0.000000
SEMIO_LUST                0.000000
SEMIO_ERL                 0.000000
SEMIO_KULT                0.000000
SEMIO_RAT                 0.000000
SEMIO_KRIT                0.000000
SEMIO_DOM                 0.000000
SEMIO_TRADV               0.000000
ZABEOTYP                  0.000000
dtype: float64
In [1948]:
missing_sorted = missing.sort_values(ascending=False)
In [1950]:
missing_sorted.index
Out[1950]:
Index(['TITEL_KZ', 'AGER_TYP', 'KK_KUNDENTYP', 'KBA05_BAUMAX', 'GEBURTSJAHR',
       'ALTER_HH', 'KKK', 'REGIOTYP', 'W_KEIT_KIND_HH', 'MOBI_REGIO',
       'KBA05_GBZ', 'KBA05_ANTG4', 'KBA05_ANTG2', 'KBA05_ANTG1', 'KBA05_ANTG3',
       'PLZ8_ANTG3', 'PLZ8_ANTG2', 'PLZ8_GBZ', 'PLZ8_HHZ', 'PLZ8_ANTG1',
       'PLZ8_BAUMAX', 'PLZ8_ANTG4', 'SHOPPER_TYP', 'VERS_TYP', 'HEALTH_TYP',
       'NATIONALITAET_KZ', 'PRAEGENDE_JUGENDJAHRE', 'KBA13_ANZAHL_PKW',
       'ANZ_HAUSHALTE_AKTIV', 'CAMEO_INTL_2015', 'CAMEO_DEU_2015',
       'CAMEO_DEUG_2015', 'LP_LEBENSPHASE_FEIN', 'RELAT_AB', 'ARBEIT',
       'ORTSGR_KLS9', 'ANZ_HH_TITEL', 'LP_LEBENSPHASE_GROB', 'INNENSTADT',
       'EWDICHTE', 'BALLRAUM', 'GEBAEUDETYP_RASTER', 'MIN_GEBAEUDEJAHR',
       'GEBAEUDETYP', 'WOHNLAGE', 'OST_WEST_KZ', 'LP_FAMILIE_GROB',
       'LP_FAMILIE_FEIN', 'KONSUMNAEHE', 'WOHNDAUER_2008', 'SOHO_KZ',
       'ANZ_TITEL', 'ANZ_PERSONEN', 'HH_EINKOMMEN_SCORE', 'LP_STATUS_GROB',
       'ONLINE_AFFINITAET', 'RETOURTYP_BK_S', 'LP_STATUS_FEIN',
       'CJT_GESAMTTYP', 'GFK_URLAUBERTYP', 'ALTERSKATEGORIE_GROB', 'SEMIO_FAM',
       'SEMIO_TRADV', 'SEMIO_DOM', 'SEMIO_KRIT', 'SEMIO_RAT', 'SEMIO_KULT',
       'SEMIO_ERL', 'SEMIO_LUST', 'SEMIO_VERT', 'SEMIO_MAT', 'SEMIO_REL',
       'FINANZ_SPARER', 'SEMIO_PFLICHT', 'SEMIO_SOZ', 'SEMIO_KAEM',
       'ANREDE_KZ', 'FINANZ_ANLEGER', 'FINANZ_VORSORGER', 'FINANZ_MINIMALIST',
       'GREEN_AVANTGARDE', 'FINANZ_HAUSBAUER', 'FINANZTYP',
       'FINANZ_UNAUFFAELLIGER', 'ZABEOTYP'],
      dtype='object')
In [1952]:
missing_sorted.values
Out[1952]:
array([99.75763587, 76.9554353 , 65.59674873, 53.46866827, 44.02028229,
       34.81369941, 17.73566826, 17.73566826, 16.60508449, 14.95970135,
       14.95970135, 14.95970135, 14.95970135, 14.95970135, 14.95970135,
       13.07363718, 13.07363718, 13.07363718, 13.07363718, 13.07363718,
       13.07363718, 13.07363718, 12.47681551, 12.47681551, 12.47681551,
       12.15355114, 12.13660809, 11.87135402, 11.17691347, 11.14785222,
       11.14785222, 11.14785222, 10.95485856, 10.92602172, 10.92602172,
       10.91468895, 10.88484226, 10.61150938, 10.51815431, 10.51815431,
       10.51815431, 10.45251402, 10.45172858, 10.45172858, 10.45172858,
       10.45172858,  8.72869917,  8.72869917,  8.2997371 ,  8.24700046,
        8.24700046,  8.24700046,  8.24700046,  2.05874862,  0.54464605,
        0.54464605,  0.54464605,  0.54464605,  0.54464605,  0.54464605,
        0.32326438,  0.        ,  0.        ,  0.        ,  0.        ,
        0.        ,  0.        ,  0.        ,  0.        ,  0.        ,
        0.        ,  0.        ,  0.        ,  0.        ,  0.        ,
        0.        ,  0.        ,  0.        ,  0.        ,  0.        ,
        0.        ,  0.        ,  0.        ,  0.        ,  0.        ])
In [1954]:
# Create a bar plot instead of a histogram for better representation

plt.figure(figsize=(12, 9))
plt.bar(missing_sorted.index, missing_sorted.values)
plt.xticks(rotation=75);
No description has been provided for this image
  • Interesing first 6 bars, let us investigate further
In [1957]:
# Perform an assessment of how much missing data there is in each column of the
# dataset.


plt.figure(figsize=(12, 9))
plt.hist(missing_sorted.values, bins=len(missing)+10, edgecolor='white', label=missing.index);
No description has been provided for this image

For this analysis, we drop the 6 Columns with more than 30% of its values missing, because they make up the last 6 little bars you see in the above plot¶

  • Relating data in terms of percentage, we see how TITEL_KZ especially has 99.757636% of its values missing. Attempting to fill this would equate adding noise to our dataset, as the odds of knowing how to fill it are low
  • Subsequent dimensions such as AGER_TYP and KK_KUNDENTYP columns might be dropped but only after more scrutiny as they have 76.9% and 65.6% of them empty respectfully
  • So also, KBA05_BAUMAX, GEBURTSJAHR and ALTER_HH would be dropped
  • The values of the other few columns should duely be filled with whatever method we deem fit
In [1964]:
pop = pop.drop(['TITEL_KZ','AGER_TYP','KK_KUNDENTYP','KBA05_BAUMAX','GEBURTSJAHR','ALTER_HH'], axis=1)
In [1966]:
pop.describe()
Out[1966]:
ALTERSKATEGORIE_GROB ANREDE_KZ CJT_GESAMTTYP FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER FINANZTYP GFK_URLAUBERTYP GREEN_AVANTGARDE HEALTH_TYP LP_LEBENSPHASE_FEIN LP_LEBENSPHASE_GROB LP_FAMILIE_FEIN LP_FAMILIE_GROB LP_STATUS_FEIN LP_STATUS_GROB NATIONALITAET_KZ PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV SHOPPER_TYP SOHO_KZ VERS_TYP ZABEOTYP ANZ_PERSONEN ANZ_TITEL HH_EINKOMMEN_SCORE W_KEIT_KIND_HH WOHNDAUER_2008 ANZ_HAUSHALTE_AKTIV ANZ_HH_TITEL GEBAEUDETYP KONSUMNAEHE MIN_GEBAEUDEJAHR WOHNLAGE KBA05_ANTG1 KBA05_ANTG2 KBA05_ANTG3 KBA05_ANTG4 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP KBA13_ANZAHL_PKW PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_BAUMAX PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB
count 888340.000000 891221.000000 886367.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 886367.000000 891221.000000 780025.000000 793589.000000 796649.000000 813429.000000 813429.000000 886367.000000 886367.000000 782906.000000 783057.000000 886367.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 891221.000000 780025.000000 817722.000000 780025.000000 891221.000000 817722.000000 817722.000000 872873.000000 743233.000000 817722.000000 791610.000000 794213.000000 798073.000000 817252.000000 798073.000000 798073.000000 757897.000000 757897.000000 757897.000000 757897.000000 757897.000000 797481.000000 797481.000000 797481.000000 798066.000000 733157.000000 757897.000000 886367.000000 733157.000000 785421.000000 774706.000000 774706.000000 774706.000000 774706.000000 774706.000000 774706.000000 774706.000000 793846.000000 793947.000000 793846.000000
mean 2.757217 1.522098 3.632838 3.074528 2.821039 3.401106 3.033328 2.874167 3.075121 3.790586 7.350304 0.196612 2.190129 16.332161 4.955185 3.922339 2.381976 4.791151 2.432575 1.168889 9.280709 3.419630 3.945860 4.272729 4.240609 4.001597 4.023709 4.359086 4.481405 4.025014 3.910139 4.763223 4.667550 4.445007 4.256076 3.661784 1.590134 0.008423 1.511166 3.362438 1.727637 0.004162 4.207243 4.147141 7.908791 8.354924 0.040647 2.798641 3.018452 1993.277011 4.052836 1.494277 1.265584 0.624525 0.305927 3.158580 4.153043 3.939172 4.549491 3.738306 2.723384 2.963540 2.698691 4.472086 619.701439 2.253330 2.801858 1.595426 0.699166 1.943913 3.612821 3.381087 3.166686 5.293389 3.071033
std 1.009951 0.499512 1.595021 1.321055 1.464749 1.322134 1.529603 1.486731 1.353248 1.987876 3.525723 0.397437 0.755213 12.242378 3.748974 3.941285 1.701527 3.425305 1.474315 0.475075 4.032107 1.417741 1.946564 1.915885 2.007373 1.857540 2.077746 2.022829 1.807552 1.903816 1.580306 1.830789 1.795712 1.852412 1.770137 1.707637 1.027972 0.091392 0.499876 1.352704 1.155849 0.068855 1.624057 1.784211 1.923137 15.673731 0.324028 2.656713 1.550312 3.332739 1.949539 1.403961 1.245178 1.013443 0.638725 1.329537 2.183710 1.718996 2.028919 0.923193 0.979867 1.428882 1.521524 1.836357 340.034318 0.972008 0.920309 0.986736 0.727137 1.459654 0.973967 1.111598 0.999072 2.303379 1.360532
min 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 0.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 0.000000 0.000000 1.000000 1.000000 0.000000 0.000000 1.000000 1.000000 1.000000 1.000000 0.000000 1.000000 1.000000 1985.000000 0.000000 0.000000 0.000000 0.000000 0.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000 0.000000 1.000000 0.000000 0.000000 0.000000 0.000000 0.000000 1.000000 1.000000 1.000000 1.000000 1.000000 1.000000
25% 2.000000 1.000000 2.000000 2.000000 1.000000 3.000000 2.000000 2.000000 2.000000 2.000000 5.000000 0.000000 2.000000 6.000000 2.000000 1.000000 1.000000 2.000000 1.000000 1.000000 6.000000 2.000000 2.000000 3.000000 3.000000 2.000000 2.000000 2.000000 3.000000 3.000000 3.000000 3.000000 3.000000 3.000000 3.000000 2.000000 1.000000 0.000000 1.000000 3.000000 1.000000 0.000000 3.000000 3.000000 8.000000 2.000000 0.000000 1.000000 2.000000 1992.000000 3.000000 0.000000 0.000000 0.000000 0.000000 2.000000 2.000000 2.000000 3.000000 3.000000 2.000000 2.000000 1.000000 3.000000 384.000000 1.000000 2.000000 1.000000 0.000000 1.000000 3.000000 3.000000 3.000000 4.000000 2.000000
50% 3.000000 2.000000 4.000000 3.000000 3.000000 3.000000 3.000000 3.000000 3.000000 4.000000 8.000000 0.000000 2.000000 13.000000 3.000000 1.000000 1.000000 4.000000 2.000000 1.000000 9.000000 3.000000 4.000000 4.000000 4.000000 4.000000 4.000000 5.000000 4.000000 4.000000 4.000000 5.000000 5.000000 5.000000 4.000000 3.000000 2.000000 0.000000 2.000000 3.000000 1.000000 0.000000 5.000000 4.000000 9.000000 4.000000 0.000000 1.000000 3.000000 1992.000000 3.000000 1.000000 1.000000 0.000000 0.000000 3.000000 5.000000 4.000000 5.000000 4.000000 3.000000 3.000000 3.000000 5.000000 549.000000 2.000000 3.000000 2.000000 1.000000 1.000000 4.000000 3.000000 3.000000 5.000000 3.000000
75% 4.000000 2.000000 5.000000 4.000000 4.000000 5.000000 5.000000 4.000000 4.000000 6.000000 10.000000 0.000000 3.000000 28.000000 8.000000 8.000000 4.000000 9.000000 4.000000 1.000000 14.000000 5.000000 6.000000 6.000000 6.000000 5.000000 6.000000 6.000000 6.000000 5.000000 5.000000 6.000000 6.000000 6.000000 6.000000 5.000000 2.000000 0.000000 2.000000 4.000000 2.000000 0.000000 6.000000 6.000000 9.000000 10.000000 0.000000 3.000000 4.000000 1993.000000 5.000000 3.000000 2.000000 1.000000 0.000000 4.000000 6.000000 6.000000 6.000000 4.000000 3.000000 4.000000 4.000000 6.000000 778.000000 3.000000 3.000000 2.000000 1.000000 3.000000 4.000000 4.000000 4.000000 7.000000 4.000000
max 4.000000 2.000000 6.000000 5.000000 5.000000 5.000000 5.000000 5.000000 5.000000 6.000000 12.000000 1.000000 3.000000 40.000000 12.000000 11.000000 5.000000 10.000000 5.000000 3.000000 15.000000 5.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 7.000000 3.000000 1.000000 2.000000 6.000000 45.000000 6.000000 6.000000 6.000000 9.000000 595.000000 23.000000 8.000000 7.000000 2016.000000 8.000000 4.000000 4.000000 3.000000 2.000000 5.000000 7.000000 6.000000 8.000000 5.000000 4.000000 6.000000 5.000000 7.000000 2300.000000 4.000000 4.000000 3.000000 2.000000 5.000000 5.000000 5.000000 5.000000 9.000000 5.000000

Knowing outliers occur below and over our Lower and Upper Bounds respectively, we can use BoxPlots to visualize these points

In [1969]:
pop.columns
Out[1969]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'CJT_GESAMTTYP',
       'FINANZ_MINIMALIST', 'FINANZ_SPARER', 'FINANZ_VORSORGER',
       'FINANZ_ANLEGER', 'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER',
       'FINANZTYP', 'GFK_URLAUBERTYP', 'GREEN_AVANTGARDE', 'HEALTH_TYP',
       'LP_LEBENSPHASE_FEIN', 'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN',
       'LP_FAMILIE_GROB', 'LP_STATUS_FEIN', 'LP_STATUS_GROB',
       'NATIONALITAET_KZ', 'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S',
       'SEMIO_SOZ', 'SEMIO_FAM', 'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT',
       'SEMIO_LUST', 'SEMIO_ERL', 'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT',
       'SEMIO_DOM', 'SEMIO_KAEM', 'SEMIO_PFLICHT', 'SEMIO_TRADV',
       'SHOPPER_TYP', 'SOHO_KZ', 'VERS_TYP', 'ZABEOTYP', 'ANZ_PERSONEN',
       'ANZ_TITEL', 'HH_EINKOMMEN_SCORE', 'W_KEIT_KIND_HH', 'WOHNDAUER_2008',
       'ANZ_HAUSHALTE_AKTIV', 'ANZ_HH_TITEL', 'GEBAEUDETYP', 'KONSUMNAEHE',
       'MIN_GEBAEUDEJAHR', 'OST_WEST_KZ', 'WOHNLAGE', 'CAMEO_DEUG_2015',
       'CAMEO_DEU_2015', 'CAMEO_INTL_2015', 'KBA05_ANTG1', 'KBA05_ANTG2',
       'KBA05_ANTG3', 'KBA05_ANTG4', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE',
       'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO',
       'ONLINE_AFFINITAET', 'REGIOTYP', 'KBA13_ANZAHL_PKW', 'PLZ8_ANTG1',
       'PLZ8_ANTG2', 'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_BAUMAX', 'PLZ8_HHZ',
       'PLZ8_GBZ', 'ARBEIT', 'ORTSGR_KLS9', 'RELAT_AB'],
      dtype='object')
  • After running the first time, I realised the loop broke because some columns didnt have numeric values
In [1972]:
valueless_columns = []

for col in pop.columns:
    try:
        pop[col].plot(kind='box')
        plt.title(f'Box plot for {col}')
        plt.show()
    except TypeError:
        valueless_columns.append(col)
        continue
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
In [1973]:
print(valueless_columns)

['OST_WEST_KZ', 'CAMEO_DEUG_2015', 'CAMEO_DEU_2015', 'CAMEO_INTL_2015']
In [1976]:
for value in valueless_columns:
    print(pop[value].value_counts().reset_index().iloc[:,0].values)     #checking the values that are alien to us
    print(f'Values in {value} are above this text\n\n')

['W' 'O']
Values in OST_WEST_KZ are above this text


['8' '9' '6' '4' '3' '2' '7' '5' '1']
Values in CAMEO_DEUG_2015 are above this text


['6B' '8A' '4C' '2D' '3C' '7A' '3D' '8B' '4A' '8C' '9D' '9B' '9C' '7B'
 '9A' '2C' '8D' '6E' '2B' '5D' '6C' '2A' '5A' '1D' '1A' '3A' '5B' '5C'
 '7C' '4B' '4D' '3B' '6A' '9E' '6D' '6F' '7D' '4E' '1E' '7E' '1C' '5F'
 '1B' '5E']
Values in CAMEO_DEU_2015 are above this text


['51' '41' '24' '14' '43' '54' '25' '22' '23' '13' '45' '55' '52' '31'
 '34' '15' '44' '12' '35' '32' '33']
Values in CAMEO_INTL_2015 are above this text
  • We missed these when visually confirming the output of 'check_value_counts()'
In [1979]:
pop[valueless_columns].info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 891221 entries, 0 to 891220
Data columns (total 4 columns):
 #   Column           Non-Null Count   Dtype 
---  ------           --------------   ----- 
 0   OST_WEST_KZ      798073 non-null  object
 1   CAMEO_DEUG_2015  791869 non-null  object
 2   CAMEO_DEU_2015   791869 non-null  object
 3   CAMEO_INTL_2015  791869 non-null  object
dtypes: object(4)
memory usage: 27.2+ MB
  • From this, we can see how the CAMEO_INTL_2015, CAMEO_DEUG_2015 only need to be encoded properly.
In [1982]:
pop['CAMEO_INTL_2015'] = pd.to_numeric(pop['CAMEO_INTL_2015'], downcast='integer', errors='coerce')    
pop['CAMEO_DEUG_2015'] = pd.to_numeric(pop['CAMEO_DEUG_2015'], downcast='integer', errors='coerce')
In [1984]:
pop.head().select_dtypes(include='number').columns
Out[1984]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'CJT_GESAMTTYP',
       'FINANZ_MINIMALIST', 'FINANZ_SPARER', 'FINANZ_VORSORGER',
       'FINANZ_ANLEGER', 'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER',
       'FINANZTYP', 'GFK_URLAUBERTYP', 'GREEN_AVANTGARDE', 'HEALTH_TYP',
       'LP_LEBENSPHASE_FEIN', 'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN',
       'LP_FAMILIE_GROB', 'LP_STATUS_FEIN', 'LP_STATUS_GROB',
       'NATIONALITAET_KZ', 'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S',
       'SEMIO_SOZ', 'SEMIO_FAM', 'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT',
       'SEMIO_LUST', 'SEMIO_ERL', 'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT',
       'SEMIO_DOM', 'SEMIO_KAEM', 'SEMIO_PFLICHT', 'SEMIO_TRADV',
       'SHOPPER_TYP', 'SOHO_KZ', 'VERS_TYP', 'ZABEOTYP', 'ANZ_PERSONEN',
       'ANZ_TITEL', 'HH_EINKOMMEN_SCORE', 'W_KEIT_KIND_HH', 'WOHNDAUER_2008',
       'ANZ_HAUSHALTE_AKTIV', 'ANZ_HH_TITEL', 'GEBAEUDETYP', 'KONSUMNAEHE',
       'MIN_GEBAEUDEJAHR', 'WOHNLAGE', 'CAMEO_DEUG_2015', 'CAMEO_INTL_2015',
       'KBA05_ANTG1', 'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_ANTG4', 'KBA05_GBZ',
       'BALLRAUM', 'EWDICHTE', 'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK',
       'MOBI_REGIO', 'ONLINE_AFFINITAET', 'REGIOTYP', 'KBA13_ANZAHL_PKW',
       'PLZ8_ANTG1', 'PLZ8_ANTG2', 'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_BAUMAX',
       'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT', 'ORTSGR_KLS9', 'RELAT_AB'],
      dtype='object')
In [1986]:
pop.head().select_dtypes(include='object').columns    #verifying what our are as they should
Out[1986]:
Index(['OST_WEST_KZ', 'CAMEO_DEU_2015'], dtype='object')
In [1988]:
len(pop.head().select_dtypes(include='number').columns)
Out[1988]:
77

Dropping Columns now¶

  • On first thought, I thought the CAMEO_DEU_2015 itself needs an extraction of the Alphabets begind each letter, but iinvestigating further, I realise we dont need to decode its values
In [1992]:
pop = pop.drop(['CAMEO_DEU_2015'], axis=1)
  • Knowing well that Z score being a function of how a point deviates from other points, we can use a high Z threshold to remove outliers, but not too many as to alter out analysis

​

In [1995]:
# Investigate patterns in the amount of missing data in each column.
In [1997]:
pop.head().select_dtypes(include='number').columns
Out[1997]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'CJT_GESAMTTYP',
       'FINANZ_MINIMALIST', 'FINANZ_SPARER', 'FINANZ_VORSORGER',
       'FINANZ_ANLEGER', 'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER',
       'FINANZTYP', 'GFK_URLAUBERTYP', 'GREEN_AVANTGARDE', 'HEALTH_TYP',
       'LP_LEBENSPHASE_FEIN', 'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN',
       'LP_FAMILIE_GROB', 'LP_STATUS_FEIN', 'LP_STATUS_GROB',
       'NATIONALITAET_KZ', 'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S',
       'SEMIO_SOZ', 'SEMIO_FAM', 'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT',
       'SEMIO_LUST', 'SEMIO_ERL', 'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT',
       'SEMIO_DOM', 'SEMIO_KAEM', 'SEMIO_PFLICHT', 'SEMIO_TRADV',
       'SHOPPER_TYP', 'SOHO_KZ', 'VERS_TYP', 'ZABEOTYP', 'ANZ_PERSONEN',
       'ANZ_TITEL', 'HH_EINKOMMEN_SCORE', 'W_KEIT_KIND_HH', 'WOHNDAUER_2008',
       'ANZ_HAUSHALTE_AKTIV', 'ANZ_HH_TITEL', 'GEBAEUDETYP', 'KONSUMNAEHE',
       'MIN_GEBAEUDEJAHR', 'WOHNLAGE', 'CAMEO_DEUG_2015', 'CAMEO_INTL_2015',
       'KBA05_ANTG1', 'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_ANTG4', 'KBA05_GBZ',
       'BALLRAUM', 'EWDICHTE', 'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK',
       'MOBI_REGIO', 'ONLINE_AFFINITAET', 'REGIOTYP', 'KBA13_ANZAHL_PKW',
       'PLZ8_ANTG1', 'PLZ8_ANTG2', 'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_BAUMAX',
       'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT', 'ORTSGR_KLS9', 'RELAT_AB'],
      dtype='object')
In [1999]:
pop.select_dtypes(include='object').keys()
Out[1999]:
Index(['OST_WEST_KZ'], dtype='object')
In [2001]:
pop.select_dtypes(include='number').keys()
Out[2001]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'CJT_GESAMTTYP',
       'FINANZ_MINIMALIST', 'FINANZ_SPARER', 'FINANZ_VORSORGER',
       'FINANZ_ANLEGER', 'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER',
       'FINANZTYP', 'GFK_URLAUBERTYP', 'GREEN_AVANTGARDE', 'HEALTH_TYP',
       'LP_LEBENSPHASE_FEIN', 'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN',
       'LP_FAMILIE_GROB', 'LP_STATUS_FEIN', 'LP_STATUS_GROB',
       'NATIONALITAET_KZ', 'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S',
       'SEMIO_SOZ', 'SEMIO_FAM', 'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT',
       'SEMIO_LUST', 'SEMIO_ERL', 'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT',
       'SEMIO_DOM', 'SEMIO_KAEM', 'SEMIO_PFLICHT', 'SEMIO_TRADV',
       'SHOPPER_TYP', 'SOHO_KZ', 'VERS_TYP', 'ZABEOTYP', 'ANZ_PERSONEN',
       'ANZ_TITEL', 'HH_EINKOMMEN_SCORE', 'W_KEIT_KIND_HH', 'WOHNDAUER_2008',
       'ANZ_HAUSHALTE_AKTIV', 'ANZ_HH_TITEL', 'GEBAEUDETYP', 'KONSUMNAEHE',
       'MIN_GEBAEUDEJAHR', 'WOHNLAGE', 'CAMEO_DEUG_2015', 'CAMEO_INTL_2015',
       'KBA05_ANTG1', 'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_ANTG4', 'KBA05_GBZ',
       'BALLRAUM', 'EWDICHTE', 'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK',
       'MOBI_REGIO', 'ONLINE_AFFINITAET', 'REGIOTYP', 'KBA13_ANZAHL_PKW',
       'PLZ8_ANTG1', 'PLZ8_ANTG2', 'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_BAUMAX',
       'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT', 'ORTSGR_KLS9', 'RELAT_AB'],
      dtype='object')
In [2003]:
pop['OST_WEST_KZ'].value_counts()
Out[2003]:
OST_WEST_KZ
W    629528
O    168545
Name: count, dtype: int64
In [2005]:
pop['OST_WEST_KZ'] = pop['OST_WEST_KZ'].replace({'W':0, 'O':1})    #numeric encoding
In [2007]:
pop['OST_WEST_KZ'] = pd.to_numeric(pop['OST_WEST_KZ'], errors='coerce')       #converting to Int, creating Nans to be worked on
In [2009]:
pop['OST_WEST_KZ'].value_counts()
Out[2009]:
OST_WEST_KZ
0.0    629528
1.0    168545
Name: count, dtype: int64
In [2011]:
variances = pop.var()

plt.figure(figsize=(10, 6))
plt.hist(variances, bins=10, color='blue', alpha=0.7)
plt.title('Histogram of Feature Variances')
plt.xlabel('Variance')
plt.ylabel('Frequency')
plt.grid(axis='y', alpha=0.75)
plt.axvline(x=variances.mean(), color='red', linestyle='--', label='Mean Variance')
plt.legend()
plt.show()
No description has been provided for this image
In [2013]:
variance_count = pop.var().sort_values(ascending=False)     #computes the variance for a dataframe
variance_count
Out[2013]:
KBA13_ANZAHL_PKW         115623.337229
ANZ_HAUSHALTE_AKTIV         245.665852
CAMEO_INTL_2015             196.854351
LP_LEBENSPHASE_FEIN         149.875812
PRAEGENDE_JUGENDJAHRE        16.257885
LP_FAMILIE_FEIN              15.533731
LP_LEBENSPHASE_GROB          14.054807
GFK_URLAUBERTYP              12.430724
LP_STATUS_FEIN               11.732715
MIN_GEBAEUDEJAHR             11.107149
GEBAEUDETYP                   7.058126
CAMEO_DEUG_2015               6.199253
ORTSGR_KLS9                   5.305553
BALLRAUM                      4.768589
SEMIO_VERT                    4.317029
INNENSTADT                    4.116513
SEMIO_LUST                    4.091839
SEMIO_REL                     4.029545
FINANZTYP                     3.951651
WOHNLAGE                      3.800701
SEMIO_SOZ                     3.789112
WOHNDAUER_2008                3.698456
SEMIO_FAM                     3.670615
SEMIO_KULT                    3.624516
SEMIO_MAT                     3.450455
SEMIO_KAEM                    3.431431
REGIOTYP                      3.372207
SEMIO_KRIT                    3.351790
SEMIO_ERL                     3.267244
SEMIO_DOM                     3.224582
W_KEIT_KIND_HH                3.183411
SEMIO_PFLICHT                 3.133384
EWDICHTE                      2.954946
SEMIO_TRADV                   2.916023
LP_FAMILIE_GROB               2.895195
HH_EINKOMMEN_SCORE            2.637561
CJT_GESAMTTYP                 2.544092
SEMIO_RAT                     2.497367
KONSUMNAEHE                   2.403467
FINANZ_ANLEGER                2.339684
ONLINE_AFFINITAET             2.315034
FINANZ_UNAUFFAELLIGER         2.210369
LP_STATUS_GROB                2.173603
FINANZ_SPARER                 2.145489
PLZ8_BAUMAX                   2.130589
MOBI_REGIO                    2.041704
RETOURTYP_BK_S                2.009990
KBA05_ANTG1                   1.971106
RELAT_AB                      1.851048
FINANZ_HAUSBAUER              1.831280
ZABEOTYP                      1.829809
KBA05_GBZ                     1.767670
FINANZ_VORSORGER              1.748039
FINANZ_MINIMALIST             1.745186
KBA05_ANTG2                   1.550467
ANZ_PERSONEN                  1.335988
PLZ8_GBZ                      1.235649
SHOPPER_TYP                   1.056726
KBA05_ANTG3                   1.027067
ALTERSKATEGORIE_GROB          1.020000
ARBEIT                        0.998145
PLZ8_ANTG3                    0.973648
KKK                           0.960140
PLZ8_HHZ                      0.948611
PLZ8_ANTG1                    0.944799
GEBAEUDETYP_RASTER            0.852285
PLZ8_ANTG2                    0.846969
HEALTH_TYP                    0.570347
PLZ8_ANTG4                    0.528728
KBA05_ANTG4                   0.407970
VERS_TYP                      0.249876
ANREDE_KZ                     0.249512
NATIONALITAET_KZ              0.225696
OST_WEST_KZ                   0.166589
GREEN_AVANTGARDE              0.157956
ANZ_HH_TITEL                  0.104994
SOHO_KZ                       0.008352
ANZ_TITEL                     0.004741
dtype: float64
In [2015]:
variance_count.iloc[0]
Out[2015]:
115623.33722931633
In [2017]:
variances = pop.var()

plt.figure(figsize=(10, 6))
plt.hist(variances, bins=10, color='blue', alpha=0.7)
plt.title('Histogram of Feature Variances')
plt.xlabel('Variance')
plt.ylabel('Frequency')
plt.grid(axis='both', alpha=0.75)
plt.axvline(x=variances.mean(), color='red', linestyle='--', label='Mean Variance')
plt.legend()
plt.show()
No description has been provided for this image
  • Having a variance of 115623.337229, KBA13_ANZAHL_PKW, would totally skew our entire data, it is only logical we take it out really.
  • ANZ_HH_TITEL, SOHO_KZ, ANZ_TITEL columns have 0.104994, 0.008352 and 0.004741 variability in total, very minimal, we can drop these too
In [2020]:
variances_to_drop = ['KBA13_ANZAHL_PKW']
In [2022]:
pop = pop.drop(axis=1, columns=variances_to_drop)
In [2024]:
def identify_outlier_columns(df: pd.DataFrame, threshold: int=4) -> list:
    """
    INPUT:
    df: pandas dataframe to identify outlier columns
    threshold: integer value for valyes that fall outside mean +- 3* STD

    OUTPUT:
    outlier_columns: a list stating what columns are outliers
    """
    outlier_columns = []
    for column in df.select_dtypes(include=[np.number]).columns:
        Q1 = df[column].quantile(0.25)
        Q3 = df[column].quantile(0.75)
        IQR = Q3 - Q1
        lower_bound = Q1 - (threshold * IQR)
        upper_bound = Q3 + (threshold * IQR)
        if df[(df[column] < lower_bound) | (df[column] > upper_bound)].shape[0] > 0:
            outlier_columns.append(column)
    return outlier_columns
In [2026]:
pop_outliers = identify_outlier_columns(pop)
pop_outliers
Out[2026]:
['GREEN_AVANTGARDE',
 'NATIONALITAET_KZ',
 'SOHO_KZ',
 'ANZ_PERSONEN',
 'ANZ_TITEL',
 'WOHNDAUER_2008',
 'ANZ_HAUSHALTE_AKTIV',
 'ANZ_HH_TITEL',
 'MIN_GEBAEUDEJAHR',
 'OST_WEST_KZ',
 'KBA05_ANTG4']
In [2028]:
pop = pop.drop(columns=pop_outliers, axis=1)
In [2030]:
pop.shape[1]
Out[2030]:
66
In [2032]:
# Remove the outlier columns from the dataset. (You'll perform other data
# engineering tasks such as re-encoding and imputation later.)

Discussion 1.1.2: Assess Missing / Outlier Data in Each Column¶

Although these were already discussed in earlier markdown cells, underneath is the summary

  • TITEL_KZ especially has 99.757636% of its values missing.
  • Columns AGER_TYP and KK_KUNDENTYP are taken under scrutiny as they have 76.9% and 65.6% of them empty respectfully.
  • OST_WEST_KZ, CAMEO_DEUG_2015, CAMEO_DEU_2015, CAMEO_INTL_2015 columns specicially were filled with unformattable data of different types, some we couldn't work with, the others we could wrangle.
  • In the end, OST_WEST_KZ, TITEL_KZ, `CAMEO_DEU_2015' had to be dropped after the above considerations
  • KBA13_ANZAHL_PKW, ANZ_HH_TITEL, SOHO_KZ, ANZ_TITEL have also been dropped because of their variance values
  • GREEN_AVANTGARDE,NATIONALITAET_KZ,ANZ_PERSONEN,WOHNDAUER_2008,ANZ_HAUSHALTE_AKTIV,MIN_GEBAEUDEJAHR,KBA05_ANTG4 didnt meet our threshold for outliers

Step 1.1.3: Assess Missing Data in Each Row¶

Now, you'll perform a similar assessment for the rows of the dataset. How much data is missing in each row? As with the columns, you should see some groups of points that have a very different numbers of missing values. Divide the data into two subsets: one for data points that are above some threshold for missing values, and a second subset for points below that threshold.

In order to know what to do with the outlier rows, we should see if the distribution of data values on columns that are not missing data (or are missing very little data) are similar or different between the two groups. Select at least five of these columns and compare the distribution of values.

  • You can use seaborn's countplot() function to create a bar chart of code frequencies and matplotlib's subplot() function to put bar charts for the two subplots side by side.
  • To reduce repeated code, you might want to write a function that can perform this comparison, taking as one of its arguments a column to be compared.

Depending on what you observe in your comparison, this will have implications on how you approach your conclusions later in the analysis. If the distributions of non-missing features look similar between the data with many missing values and the data with few or no missing values, then we could argue that simply dropping those points from the analysis won't present a major issue. On the other hand, if the data with many missing values looks very different from the data with few or no missing values, then we should make a note on those data as special. We'll revisit these data later on. Either way, you should continue your analysis for now using just the subset of the data with few or no missing values.

In [2036]:
# How much data is missing in each row of the dataset?

missing_rows = ((pop.isna().sum(axis=1)/pop.shape[1])*100)   #displayingg what percentage of each row is missing
In [2038]:
plt.hist(missing_rows);
plt.xlabel('Percentage of missing values')
plt.ylabel('Number of Rows')
plt.title('Row Nan Plot');
No description has been provided for this image
  • From the plot, a 30% threshold is fine for a segregation really
In [2041]:
pop.iloc[0].reset_index().T   #examining our first row
Out[2041]:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
index ALTERSKATEGORIE_GROB ANREDE_KZ CJT_GESAMTTYP FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER FINANZTYP GFK_URLAUBERTYP HEALTH_TYP LP_LEBENSPHASE_FEIN LP_LEBENSPHASE_GROB LP_FAMILIE_FEIN LP_FAMILIE_GROB LP_STATUS_FEIN LP_STATUS_GROB PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV SHOPPER_TYP VERS_TYP ZABEOTYP HH_EINKOMMEN_SCORE W_KEIT_KIND_HH GEBAEUDETYP KONSUMNAEHE WOHNLAGE CAMEO_DEUG_2015 CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_ANTG3 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_BAUMAX PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB
0 2.0 1.0 2.0 3.0 4.0 3.0 5.0 5.0 3.0 4.0 10.0 NaN 15.0 4.0 2.0 2.0 1.0 1.0 NaN 5.0 2.0 6.0 7.0 5.0 1.0 5.0 3.0 3.0 4.0 7.0 6.0 6.0 5.0 3.0 NaN NaN 3.0 2.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
In [2043]:
len(missing_rows)
Out[2043]:
891221
In [2045]:
excess_nans = pop[missing_rows >= 30]                 #rows with more than 30% of their values missing
tolerable_nans = pop[missing_rows < 30]           #rows with at least 30% of their values intact
In [2047]:
print(len(excess_nans))
print(len(tolerable_nans))

93557
797664
In [2049]:
row_thresh = (len(excess_nans) / (len(excess_nans)+len(tolerable_nans)))*100
In [2051]:
print(f'{round(row_thresh)}% of our rows dont meet our 30 percent threshold')

10% of our rows dont meet our 30 percent threshold
In [2053]:
pop.head()
Out[2053]:
ALTERSKATEGORIE_GROB ANREDE_KZ CJT_GESAMTTYP FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER FINANZTYP GFK_URLAUBERTYP HEALTH_TYP LP_LEBENSPHASE_FEIN LP_LEBENSPHASE_GROB LP_FAMILIE_FEIN LP_FAMILIE_GROB LP_STATUS_FEIN LP_STATUS_GROB PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV SHOPPER_TYP VERS_TYP ZABEOTYP HH_EINKOMMEN_SCORE W_KEIT_KIND_HH GEBAEUDETYP KONSUMNAEHE WOHNLAGE CAMEO_DEUG_2015 CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_ANTG3 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_BAUMAX PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB
0 2.0 1 2.0 3 4 3 5 5 3 4 10.0 NaN 15.0 4.0 2.0 2.0 1.0 1.0 NaN 5.0 2 6 7 5 1 5 3 3 4 7 6 6 5 3 NaN NaN 3 2.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
1 1.0 2 5.0 1 5 2 5 4 5 1 10.0 3.0 21.0 6.0 5.0 3.0 2.0 1.0 14.0 1.0 5 4 4 3 1 2 2 3 6 4 7 4 7 6 3.0 2.0 5 6.0 3.0 8.0 1.0 4.0 8.0 51.0 0.0 0.0 0.0 1.0 6.0 3.0 8.0 3.0 2.0 1.0 3.0 3.0 2.0 3.0 2.0 1.0 1.0 5.0 4.0 3.0 5.0 4.0
2 3.0 2 3.0 1 4 1 2 3 5 1 10.0 3.0 3.0 1.0 1.0 1.0 3.0 2.0 15.0 3.0 4 1 3 3 4 4 6 3 4 7 7 7 3 3 2.0 1.0 5 4.0 3.0 1.0 5.0 2.0 4.0 24.0 1.0 3.0 1.0 3.0 2.0 4.0 4.0 4.0 2.0 3.0 2.0 2.0 3.0 3.0 1.0 0.0 1.0 4.0 4.0 3.0 5.0 2.0
3 4.0 2 2.0 4 2 5 2 1 2 6 1.0 2.0 NaN NaN NaN NaN 9.0 4.0 8.0 2.0 5 1 2 1 4 4 7 4 3 4 4 5 4 4 1.0 1.0 3 1.0 NaN 1.0 4.0 7.0 2.0 12.0 4.0 1.0 0.0 4.0 4.0 2.0 6.0 4.0 NaN 4.0 1.0 NaN 2.0 2.0 2.0 0.0 1.0 3.0 4.0 2.0 3.0 3.0
4 3.0 1 5.0 4 3 4 1 3 2 5 5.0 3.0 32.0 10.0 10.0 5.0 3.0 2.0 8.0 5.0 6 4 4 2 7 4 4 6 2 3 2 2 4 2 2.0 2.0 4 5.0 2.0 1.0 4.0 3.0 6.0 43.0 1.0 4.0 1.0 3.0 2.0 5.0 1.0 5.0 3.0 3.0 5.0 5.0 2.0 4.0 2.0 1.0 2.0 3.0 3.0 4.0 6.0 5.0
In [2055]:
pop.shape
Out[2055]:
(891221, 66)
In [2057]:
columns_with_few_nans = pop.columns[pop.isna().mean() < 0.2].tolist()[:11]
In [2059]:
columns_with_few_nans
Out[2059]:
['ALTERSKATEGORIE_GROB',
 'ANREDE_KZ',
 'CJT_GESAMTTYP',
 'FINANZ_MINIMALIST',
 'FINANZ_SPARER',
 'FINANZ_VORSORGER',
 'FINANZ_ANLEGER',
 'FINANZ_UNAUFFAELLIGER',
 'FINANZ_HAUSBAUER',
 'FINANZTYP',
 'GFK_URLAUBERTYP']
In [2061]:
def plot_distribution(col:str) -> None:
    """
    INPUT: 
    Col(str): column in pandas dataframe
    OUTPUT:
    None: only displays column rows
    
    """
    fig, axes = plt.subplots(1, 2, figsize=(12, 6),sharex=True)

    plt.subplot(1, 2, 1)
    sns.countplot(x=col, data=tolerable_nans)
    axes[0].set_title(f'Low Amount of NaNs - {col}')

    plt.subplot(1, 2, 2)
    sns.countplot(x=col, data=excess_nans)
    axes[1].set_title(f'High Amount of Nans - {col}')
    
    plt.show()
In [2063]:
for col in columns_with_few_nans:
    plot_distribution(col)
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
  • The difference between our tolerable nans and excess nans between these 5 columns with little empty values is very distinct.
In [2065]:
pop.keys()
Out[2065]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'CJT_GESAMTTYP',
       'FINANZ_MINIMALIST', 'FINANZ_SPARER', 'FINANZ_VORSORGER',
       'FINANZ_ANLEGER', 'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER',
       'FINANZTYP', 'GFK_URLAUBERTYP', 'HEALTH_TYP', 'LP_LEBENSPHASE_FEIN',
       'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN', 'LP_FAMILIE_GROB',
       'LP_STATUS_FEIN', 'LP_STATUS_GROB', 'PRAEGENDE_JUGENDJAHRE',
       'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM', 'SEMIO_REL', 'SEMIO_MAT',
       'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL', 'SEMIO_KULT', 'SEMIO_RAT',
       'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM', 'SEMIO_PFLICHT', 'SEMIO_TRADV',
       'SHOPPER_TYP', 'VERS_TYP', 'ZABEOTYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'GEBAEUDETYP', 'KONSUMNAEHE', 'WOHNLAGE',
       'CAMEO_DEUG_2015', 'CAMEO_INTL_2015', 'KBA05_ANTG1', 'KBA05_ANTG2',
       'KBA05_ANTG3', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE', 'INNENSTADT',
       'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO', 'ONLINE_AFFINITAET',
       'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2', 'PLZ8_ANTG3', 'PLZ8_ANTG4',
       'PLZ8_BAUMAX', 'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT', 'ORTSGR_KLS9',
       'RELAT_AB'],
      dtype='object')

Discussion 1.1.3: Assess Missing Data in Each Row¶

  • The dimension Anrede_Kz, a little of Alterskategorie_grob show a similarity in how their tolerable nans are compared to their excesses
  • In most cases though, the data with high amount of missing dataa is very different from the ones without missing data. THis means, using data with few or no missing values will cause incorrect predicitions.

Step 1.2: Select and Re-Encode Features¶

Checking for missing data isn't the only way in which you can prepare a dataset for analysis. Since the unsupervised learning techniques to be used will only work on data that is encoded numerically, you need to make a few encoding changes or additional assumptions to be able to make progress. In addition, while almost all of the values in the dataset are encoded using numbers, not all of them represent numeric values. Check the third column of the feature summary (feat_info) for a summary of types of measurement.

  • For numeric and interval data, these features can be kept without changes.
  • Most of the variables in the dataset are ordinal in nature. While ordinal values may technically be non-linear in spacing, make the simplifying assumption that the ordinal variables can be treated as being interval in nature (that is, kept without any changes).
  • Special handling may be necessary for the remaining two variable types: categorical, and 'mixed'.

In the first two parts of this sub-step, you will perform an investigation of the categorical and mixed-type features and make a decision on each of them, whether you will keep, drop, or re-encode each. Then, in the last part, you will create a new data frame with only the selected and engineered columns.

Data wrangling is often the trickiest part of the data analysis process, and there's a lot of it to be done here. But stick with it: once you're done with this step, you'll be ready to get to the machine learning parts of the project!

In [2070]:
# How many features are there of each data type?

feat_info['type'].value_counts()
Out[2070]:
type
ordinal        49
categorical    21
numeric         7
mixed           7
interval        1
Name: count, dtype: int64

• Leaving numeric, ordinal and interval data as they are, we are left to wrangle the: categorical and mixed data types

In [2073]:
feat_info.query('type=="categorical" or type=="mixed"')     #finding the categorical columns
Out[2073]:
attribute information_level type missing_or_unknown
0 AGER_TYP person categorical [-1,0]
2 ANREDE_KZ person categorical [-1,0]
3 CJT_GESAMTTYP person categorical [0]
10 FINANZTYP person categorical [-1]
12 GFK_URLAUBERTYP person categorical []
13 GREEN_AVANTGARDE person categorical []
15 LP_LEBENSPHASE_FEIN person mixed [0]
16 LP_LEBENSPHASE_GROB person mixed [0]
17 LP_FAMILIE_FEIN person categorical [0]
18 LP_FAMILIE_GROB person categorical [0]
19 LP_STATUS_FEIN person categorical [0]
20 LP_STATUS_GROB person categorical [0]
21 NATIONALITAET_KZ person categorical [-1,0]
22 PRAEGENDE_JUGENDJAHRE person mixed [-1,0]
38 SHOPPER_TYP person categorical [-1]
39 SOHO_KZ person categorical [-1]
40 TITEL_KZ person categorical [-1,0]
41 VERS_TYP person categorical [-1]
42 ZABEOTYP person categorical [-1,9]
47 KK_KUNDENTYP household categorical [-1]
52 GEBAEUDETYP building categorical [-1,0]
55 OST_WEST_KZ building categorical [-1]
56 WOHNLAGE building mixed [-1]
57 CAMEO_DEUG_2015 microcell_rr4 categorical [-1,X]
58 CAMEO_DEU_2015 microcell_rr4 categorical [XX]
59 CAMEO_INTL_2015 microcell_rr4 mixed [-1,XX]
64 KBA05_BAUMAX microcell_rr3 mixed [-1,0]
79 PLZ8_BAUMAX macrocell_plz8 mixed [-1,0]
In [2075]:
cat_dimensions = feat_info.query('type=="categorical" or type=="mixed"')['attribute']     #finding these categorical columns
cat_dimensions
Out[2075]:
0                  AGER_TYP
2                 ANREDE_KZ
3             CJT_GESAMTTYP
10                FINANZTYP
12          GFK_URLAUBERTYP
13         GREEN_AVANTGARDE
15      LP_LEBENSPHASE_FEIN
16      LP_LEBENSPHASE_GROB
17          LP_FAMILIE_FEIN
18          LP_FAMILIE_GROB
19           LP_STATUS_FEIN
20           LP_STATUS_GROB
21         NATIONALITAET_KZ
22    PRAEGENDE_JUGENDJAHRE
38              SHOPPER_TYP
39                  SOHO_KZ
40                 TITEL_KZ
41                 VERS_TYP
42                 ZABEOTYP
47             KK_KUNDENTYP
52              GEBAEUDETYP
55              OST_WEST_KZ
56                 WOHNLAGE
57          CAMEO_DEUG_2015
58           CAMEO_DEU_2015
59          CAMEO_INTL_2015
64             KBA05_BAUMAX
79              PLZ8_BAUMAX
Name: attribute, dtype: object
In [2077]:
pop.head()
Out[2077]:
ALTERSKATEGORIE_GROB ANREDE_KZ CJT_GESAMTTYP FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER FINANZTYP GFK_URLAUBERTYP HEALTH_TYP LP_LEBENSPHASE_FEIN LP_LEBENSPHASE_GROB LP_FAMILIE_FEIN LP_FAMILIE_GROB LP_STATUS_FEIN LP_STATUS_GROB PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV SHOPPER_TYP VERS_TYP ZABEOTYP HH_EINKOMMEN_SCORE W_KEIT_KIND_HH GEBAEUDETYP KONSUMNAEHE WOHNLAGE CAMEO_DEUG_2015 CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_ANTG3 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_BAUMAX PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB
0 2.0 1 2.0 3 4 3 5 5 3 4 10.0 NaN 15.0 4.0 2.0 2.0 1.0 1.0 NaN 5.0 2 6 7 5 1 5 3 3 4 7 6 6 5 3 NaN NaN 3 2.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
1 1.0 2 5.0 1 5 2 5 4 5 1 10.0 3.0 21.0 6.0 5.0 3.0 2.0 1.0 14.0 1.0 5 4 4 3 1 2 2 3 6 4 7 4 7 6 3.0 2.0 5 6.0 3.0 8.0 1.0 4.0 8.0 51.0 0.0 0.0 0.0 1.0 6.0 3.0 8.0 3.0 2.0 1.0 3.0 3.0 2.0 3.0 2.0 1.0 1.0 5.0 4.0 3.0 5.0 4.0
2 3.0 2 3.0 1 4 1 2 3 5 1 10.0 3.0 3.0 1.0 1.0 1.0 3.0 2.0 15.0 3.0 4 1 3 3 4 4 6 3 4 7 7 7 3 3 2.0 1.0 5 4.0 3.0 1.0 5.0 2.0 4.0 24.0 1.0 3.0 1.0 3.0 2.0 4.0 4.0 4.0 2.0 3.0 2.0 2.0 3.0 3.0 1.0 0.0 1.0 4.0 4.0 3.0 5.0 2.0
3 4.0 2 2.0 4 2 5 2 1 2 6 1.0 2.0 NaN NaN NaN NaN 9.0 4.0 8.0 2.0 5 1 2 1 4 4 7 4 3 4 4 5 4 4 1.0 1.0 3 1.0 NaN 1.0 4.0 7.0 2.0 12.0 4.0 1.0 0.0 4.0 4.0 2.0 6.0 4.0 NaN 4.0 1.0 NaN 2.0 2.0 2.0 0.0 1.0 3.0 4.0 2.0 3.0 3.0
4 3.0 1 5.0 4 3 4 1 3 2 5 5.0 3.0 32.0 10.0 10.0 5.0 3.0 2.0 8.0 5.0 6 4 4 2 7 4 4 6 2 3 2 2 4 2 2.0 2.0 4 5.0 2.0 1.0 4.0 3.0 6.0 43.0 1.0 4.0 1.0 3.0 2.0 5.0 1.0 5.0 3.0 3.0 5.0 5.0 2.0 4.0 2.0 1.0 2.0 3.0 3.0 4.0 6.0 5.0
In [2079]:
#pop['AGER_TYP'].value_counts().reset_index().iloc[:,0].values
In [2081]:
for col in cat_dimensions:
    try: 
        counte = pop[col].value_counts().reset_index().iloc[:,0].values
        print(f'{col} has: {counte}')
    except KeyError:
        continue

ANREDE_KZ has: [2 1]
CJT_GESAMTTYP has: [4. 3. 6. 2. 5. 1.]
FINANZTYP has: [6 1 4 2 5 3]
GFK_URLAUBERTYP has: [12.  5. 10.  8. 11.  4.  9.  3.  1.  2.  7.  6.]
LP_LEBENSPHASE_FEIN has: [ 1.  5.  6.  2.  8. 11. 29.  7. 13. 10. 31. 12. 30. 15.  3. 19. 37.  4.
 14. 20. 32. 39. 40. 27. 16. 38. 35. 34.  9. 21. 28. 24. 36. 25. 23. 22.
 18. 33. 17. 26.]
LP_LEBENSPHASE_GROB has: [ 2.  1.  3. 12.  4.  5.  9. 10. 11.  8.  6.  7.]
LP_FAMILIE_FEIN has: [ 1. 10.  2. 11.  8.  7.  4.  5.  9.  6.  3.]
LP_FAMILIE_GROB has: [1. 5. 2. 4. 3.]
LP_STATUS_FEIN has: [ 1.  9.  2. 10.  4.  5.  3.  6.  8.  7.]
LP_STATUS_GROB has: [1. 2. 4. 5. 3.]
PRAEGENDE_JUGENDJAHRE has: [14.  8.  5. 10.  3. 15. 11.  9.  6. 12.  1.  4.  2. 13.  7.]
SHOPPER_TYP has: [1. 2. 3. 0.]
VERS_TYP has: [2. 1.]
ZABEOTYP has: [3 4 1 5 6 2]
GEBAEUDETYP has: [1. 3. 8. 2. 4. 6. 5.]
WOHNLAGE has: [3. 7. 4. 2. 5. 1. 8. 0.]
CAMEO_DEUG_2015 has: [8. 9. 6. 4. 3. 2. 7. 5. 1.]
CAMEO_INTL_2015 has: [51. 41. 24. 14. 43. 54. 25. 22. 23. 13. 45. 55. 52. 31. 34. 15. 44. 12.
 35. 32. 33.]
PLZ8_BAUMAX has: [1. 5. 2. 4. 3.]
In [2083]:
multi_col = []
for col in cat_dimensions:
    column_data = pop.get(col)
    if column_data is not None:
        counte = column_data.value_counts().reset_index().iloc[:, 0].values
        print(f'{col} has: {counte} values')
        if len(counte) > 2:
            multi_col.append(col)

ANREDE_KZ has: [2 1] values
CJT_GESAMTTYP has: [4. 3. 6. 2. 5. 1.] values
FINANZTYP has: [6 1 4 2 5 3] values
GFK_URLAUBERTYP has: [12.  5. 10.  8. 11.  4.  9.  3.  1.  2.  7.  6.] values
LP_LEBENSPHASE_FEIN has: [ 1.  5.  6.  2.  8. 11. 29.  7. 13. 10. 31. 12. 30. 15.  3. 19. 37.  4.
 14. 20. 32. 39. 40. 27. 16. 38. 35. 34.  9. 21. 28. 24. 36. 25. 23. 22.
 18. 33. 17. 26.] values
LP_LEBENSPHASE_GROB has: [ 2.  1.  3. 12.  4.  5.  9. 10. 11.  8.  6.  7.] values
LP_FAMILIE_FEIN has: [ 1. 10.  2. 11.  8.  7.  4.  5.  9.  6.  3.] values
LP_FAMILIE_GROB has: [1. 5. 2. 4. 3.] values
LP_STATUS_FEIN has: [ 1.  9.  2. 10.  4.  5.  3.  6.  8.  7.] values
LP_STATUS_GROB has: [1. 2. 4. 5. 3.] values
PRAEGENDE_JUGENDJAHRE has: [14.  8.  5. 10.  3. 15. 11.  9.  6. 12.  1.  4.  2. 13.  7.] values
SHOPPER_TYP has: [1. 2. 3. 0.] values
VERS_TYP has: [2. 1.] values
ZABEOTYP has: [3 4 1 5 6 2] values
GEBAEUDETYP has: [1. 3. 8. 2. 4. 6. 5.] values
WOHNLAGE has: [3. 7. 4. 2. 5. 1. 8. 0.] values
CAMEO_DEUG_2015 has: [8. 9. 6. 4. 3. 2. 7. 5. 1.] values
CAMEO_INTL_2015 has: [51. 41. 24. 14. 43. 54. 25. 22. 23. 13. 45. 55. 52. 31. 34. 15. 44. 12.
 35. 32. 33.] values
PLZ8_BAUMAX has: [1. 5. 2. 4. 3.] values
In [2085]:
print(multi_col)

['CJT_GESAMTTYP', 'FINANZTYP', 'GFK_URLAUBERTYP', 'LP_LEBENSPHASE_FEIN', 'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN', 'LP_FAMILIE_GROB', 'LP_STATUS_FEIN', 'LP_STATUS_GROB', 'PRAEGENDE_JUGENDJAHRE', 'SHOPPER_TYP', 'ZABEOTYP', 'GEBAEUDETYP', 'WOHNLAGE', 'CAMEO_DEUG_2015', 'CAMEO_INTL_2015', 'PLZ8_BAUMAX']

Step 1.2.1: Re-Encode Categorical Features¶

For categorical data, you would ordinarily need to encode the levels as dummy variables. Depending on the number of categories, perform one of the following:

  • For binary (two-level) categoricals that take numeric values, you can keep them without needing to do anything.
  • There is one binary variable that takes on non-numeric values. For this one, you need to re-encode the values as numbers or create a dummy variable.(We already did this earlier)
  • For multi-level categoricals (three or more values), you can choose to encode the values using multiple dummy variables (e.g. via OneHotEncoder), or (to keep things straightforward) just drop them from the analysis. As always, document your choices in the Discussion section.
  • Basically, columns in the to_dummy list are the ones that have more than 2 unique values, and are still categorical

Assess categorical variables: which are binary, which are multi-level, and which one needs to be re-encoded?

In [2089]:
valid_columns = [col for col in multi_col if col in pop.columns]      # Subset the DataFrame to only columns present in the multi_col list
In [2091]:
#both needed for later encoding

valid_columns.remove('CAMEO_INTL_2015')
valid_columns.remove('PRAEGENDE_JUGENDJAHRE')
In [2093]:
print(valid_columns)

['CJT_GESAMTTYP', 'FINANZTYP', 'GFK_URLAUBERTYP', 'LP_LEBENSPHASE_FEIN', 'LP_LEBENSPHASE_GROB', 'LP_FAMILIE_FEIN', 'LP_FAMILIE_GROB', 'LP_STATUS_FEIN', 'LP_STATUS_GROB', 'SHOPPER_TYP', 'ZABEOTYP', 'GEBAEUDETYP', 'WOHNLAGE', 'CAMEO_DEUG_2015', 'PLZ8_BAUMAX']
In [2095]:
pop = pop.drop(columns=valid_columns)
In [2097]:
# Re-encode categorical variable(s) to be kept in the analysis.

Discussion 1.2.1: Re-Encode Categorical Features¶

(Double-click this cell and replace this text with your own text, reporting your findings and decisions regarding categorical features. Which ones did you keep, which did you drop, and what engineering steps did you perform?)

In [2100]:
feat_info.query('type=="mixed"')['attribute'].to_list()
Out[2100]:
['LP_LEBENSPHASE_FEIN',
 'LP_LEBENSPHASE_GROB',
 'PRAEGENDE_JUGENDJAHRE',
 'WOHNLAGE',
 'CAMEO_INTL_2015',
 'KBA05_BAUMAX',
 'PLZ8_BAUMAX']

Step 1.2.2: Engineer Mixed-Type Features¶

There are a handful of features that are marked as "mixed" in the feature summary that require special treatment in order to be included in the analysis. There are two in particular that deserve attention; the handling of the rest are up to your own choices:

  • "PRAEGENDE_JUGENDJAHRE" combines information on three dimensions: generation by decade, movement (mainstream vs. avantgarde), and nation (east vs. west). While there aren't enough levels to disentangle east from west, you should create two new variables to capture the other two dimensions: an interval-type variable for decade, and a binary variable for movement.
  • "CAMEO_INTL_2015" combines information on two axes: wealth and life stage. Break up the two-digit codes by their 'tens'-place and 'ones'-place digits into two new ordinal variables (which, for the purposes of this project, is equivalent to just treating them as their raw numeric values).
  • If you decide to keep or engineer new features around the other mixed-type features, make sure you note your steps in the Discussion section.

Reading Data_Dictionary.md, we can understand the steps described above

In [2103]:
pop['CAMEO_INTL_2015'].value_counts().reset_index().sort_values(by='CAMEO_INTL_2015')
Out[2103]:
CAMEO_INTL_2015 count
17 12.0 13249
9 13.0 26336
3 14.0 62884
15 15.0 16974
7 22.0 33155
8 23.0 26750
2 24.0 91158
6 25.0 39628
13 31.0 19024
19 32.0 10354
20 33.0 9935
14 34.0 18524
18 35.0 10356
1 41.0 92336
4 43.0 56672
16 44.0 14820
10 45.0 26132
0 51.0 133694
12 52.0 20542
5 54.0 45391
11 55.0 23955
CAMEO_INTL_2015 DECODING¶

German CAMEO: Wealth / Life Stage Typology, mapped to international code

  • -1: unknown
  • 11: Wealthy Households - Pre-Family Couples & Singles
  • 12: Wealthy Households - Young Couples With Children
  • 13: Wealthy Households - Families With School Age Children
  • 14: Wealthy Households - Older Families & Mature Couples
  • 15: Wealthy Households - Elders In Retirement
  • 21: Prosperous Households - Pre-Family Couples & Singles
  • 22: Prosperous Households - Young Couples With Children
  • 23: Prosperous Households - Families With School Age Children
  • 24: Prosperous Households - Older Families & Mature Couples
  • 25: Prosperous Households - Elders In Retirement
  • 31: Comfortable Households - Pre-Family Couples & Singles
  • 32: Comfortable Households - Young Couples With Children
  • 33: Comfortable Households - Families With School Age Children
  • 34: Comfortable Households - Older Families & Mature Couples
  • 35: Comfortable Households - Elders In Retirement
  • 41: Less Affluent Households - Pre-Family Couples & Singles
  • 42: Less Affluent Households - Young Couples With Children
  • 43: Less Affluent Households - Families With School Age Children
  • 44: Less Affluent Households - Older Families & Mature Couples
  • 45: Less Affluent Households - Elders In Retirement
  • 51: Poorer Households - Pre-Family Couples & Singles
  • 52: Poorer Households - Young Couples With Children
  • 53: Poorer Households - Families With School Age Children
  • 54: Poorer Households - Older Families & Mature Couples
  • 55: Poorer Households - Elders In Retirement
  • XX: unknown
In [2106]:
wealth = []
life_stage = []

for i in pop['CAMEO_INTL_2015'].astype(float):
    if i//10==1.0:
        wealth.append(0)    #wealthy households
    elif i//10==2.0:
        wealth.append(1)    #prosperous households
    elif i//10==3.0:
        wealth.append(2)    #comfortable households
    elif i//10==4.0:
        wealth.append(3)    #less affluent households
    else:
        wealth.append(4)    #poorer households

    if i%10==1.0:
        life_stage.append(0)      #pre-family couple
    elif i%10==2.0:
        life_stage.append(1)      #young couple
    elif i%10==3.0:
        life_stage.append(2)      #family with school age
    elif i%10==4.0:
        life_stage.append(3)      #mature couples
    else:
        life_stage.append(4)      #elder in retirement


pop['wealth'] = wealth      #encoding wealth with floor division
pop['life_stage'] = life_stage      #encoding wealth with reminder
In [2108]:
pop.keys()
Out[2108]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'FINANZ_MINIMALIST',
       'FINANZ_SPARER', 'FINANZ_VORSORGER', 'FINANZ_ANLEGER',
       'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER', 'HEALTH_TYP',
       'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM',
       'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL',
       'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM',
       'SEMIO_PFLICHT', 'SEMIO_TRADV', 'VERS_TYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'KONSUMNAEHE', 'CAMEO_INTL_2015', 'KBA05_ANTG1',
       'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE',
       'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO',
       'ONLINE_AFFINITAET', 'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2',
       'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT',
       'ORTSGR_KLS9', 'RELAT_AB', 'wealth', 'life_stage'],
      dtype='object')
In [2110]:
pop.columns.to_list()[:5]
Out[2110]:
['ALTERSKATEGORIE_GROB',
 'ANREDE_KZ',
 'FINANZ_MINIMALIST',
 'FINANZ_SPARER',
 'FINANZ_VORSORGER']
PRAEGENDE_JUGENDJAHRE DECODING¶

Dominating movement of person's youth (avantgarde vs. mainstream; east vs. west)

  • -1: unknown
  • 0: unknown
  • 1: 40s - war years (Mainstream, E+W)
  • 2: 40s - reconstruction years (Avantgarde, E+W)
  • 3: 50s - economic miracle (Mainstream, E+W)
  • 4: 50s - milk bar / Individualisation (Avantgarde, E+W)
  • 5: 60s - economic miracle (Mainstream, E+W)
  • 6: 60s - generation 68 / student protestors (Avantgarde, W)
  • 7: 60s - opponents to the building of the Wall (Avantgarde, E)
  • 8: 70s - family orientation (Mainstream, E+W)
  • 9: 70s - peace movement (Avantgarde, E+W)
  • 10: 80s - Generation Golf (Mainstream, W)
  • 11: 80s - ecological awareness (Avantgarde, W)
  • 12: 80s - FDJ / communist party youth organisation (Mainstream, E)
  • 13: 80s - Swords into ploughshares (Avantgarde, E)
  • 14: 90s - digital media kids (Mainstream, E+W)
  • 15: 90s - ecological awareness (Avantgarde, E+W)
In [2113]:
pop['PRAEGENDE_JUGENDJAHRE'].value_counts().reset_index().sort_values(by='PRAEGENDE_JUGENDJAHRE')
Out[2113]:
PRAEGENDE_JUGENDJAHRE count
10 1.0 21282
12 2.0 7479
4 3.0 55195
11 4.0 20451
2 5.0 86416
8 6.0 25652
14 7.0 4010
1 8.0 145988
7 9.0 33570
3 10.0 85808
6 11.0 35752
9 12.0 24446
13 13.0 5764
0 14.0 188697
5 15.0 42547
In [2115]:
#generation by decade, movement (mainstream vs. avantgarde)

generation = []
movement = []

for i in pop['PRAEGENDE_JUGENDJAHRE'].astype(float):
    if i == 1 or i == 2:
        generation.append(0)         #40s 
    elif i == 3 or i == 4:
        generation.append(1)         #50s 
    elif i == 5 or i == 6 or i == 7:
        generation.append(2)         #60s 
    elif i == 8 or i == 9:
        generation.append(3)         #90s 
    elif i == 14 or i == 15:
        generation.append(4)         #80s 
    else:
        generation.append(5)

    if i == 1 or i == 3 or i == 5 or i == 8 or i == 10 or i == 12 or i == 14:
        movement.append(0)      #mainstream movement
    else:
        movement.append(1)       #avantgarde movement

pop['generation']=generation      #encoding generation
pop['movement']=movement      #encoding movement type
In [2117]:
mixed_cols = feat_info.query('type=="mixed"')['attribute'].to_list()
mixed_cols
Out[2117]:
['LP_LEBENSPHASE_FEIN',
 'LP_LEBENSPHASE_GROB',
 'PRAEGENDE_JUGENDJAHRE',
 'WOHNLAGE',
 'CAMEO_INTL_2015',
 'KBA05_BAUMAX',
 'PLZ8_BAUMAX']
In [2119]:
mixed_to_drop = []

for col in mixed_cols:
    if (col != 'PRAEGENDE_JUGENDJAHRE') and (col != 'CAMEO_INTL_2015'):
        mixed_to_drop.append(col)

mixed_to_drop
Out[2119]:
['LP_LEBENSPHASE_FEIN',
 'LP_LEBENSPHASE_GROB',
 'WOHNLAGE',
 'KBA05_BAUMAX',
 'PLZ8_BAUMAX']
In [2121]:
for col in mixed_to_drop:
    try:
        pop = pop.drop(columns=col)
    except KeyError:
        continue
In [2123]:
pop.keys()
Out[2123]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'FINANZ_MINIMALIST',
       'FINANZ_SPARER', 'FINANZ_VORSORGER', 'FINANZ_ANLEGER',
       'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER', 'HEALTH_TYP',
       'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM',
       'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL',
       'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM',
       'SEMIO_PFLICHT', 'SEMIO_TRADV', 'VERS_TYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'KONSUMNAEHE', 'CAMEO_INTL_2015', 'KBA05_ANTG1',
       'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE',
       'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO',
       'ONLINE_AFFINITAET', 'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2',
       'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT',
       'ORTSGR_KLS9', 'RELAT_AB', 'wealth', 'life_stage', 'generation',
       'movement'],
      dtype='object')
In [2125]:
pop.columns
Out[2125]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'FINANZ_MINIMALIST',
       'FINANZ_SPARER', 'FINANZ_VORSORGER', 'FINANZ_ANLEGER',
       'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER', 'HEALTH_TYP',
       'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM',
       'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL',
       'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM',
       'SEMIO_PFLICHT', 'SEMIO_TRADV', 'VERS_TYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'KONSUMNAEHE', 'CAMEO_INTL_2015', 'KBA05_ANTG1',
       'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE',
       'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO',
       'ONLINE_AFFINITAET', 'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2',
       'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT',
       'ORTSGR_KLS9', 'RELAT_AB', 'wealth', 'life_stage', 'generation',
       'movement'],
      dtype='object')
In [2127]:
pop.isna().sum().sum()
Out[2127]:
3031039
In [2129]:
pop.select_dtypes(include='number').keys()
Out[2129]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'FINANZ_MINIMALIST',
       'FINANZ_SPARER', 'FINANZ_VORSORGER', 'FINANZ_ANLEGER',
       'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER', 'HEALTH_TYP',
       'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM',
       'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL',
       'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM',
       'SEMIO_PFLICHT', 'SEMIO_TRADV', 'VERS_TYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'KONSUMNAEHE', 'CAMEO_INTL_2015', 'KBA05_ANTG1',
       'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE',
       'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO',
       'ONLINE_AFFINITAET', 'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2',
       'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT',
       'ORTSGR_KLS9', 'RELAT_AB', 'wealth', 'life_stage', 'generation',
       'movement'],
      dtype='object')

Discussion 1.2.2: Engineer Mixed-Type Features¶

  • I investigated "PRAEGENDE_JUGENDJAHRE" and "CAMEO_INTL_2015" to engineer two new variables each.
  • The other dimensions, I decided tto drop them

Step 1.2.3: Complete Feature Selection¶

In order to finish this step up, you need to make sure that your data frame now only has the columns that you want to keep. To summarize, the dataframe should consist of the following:

  • All numeric, interval, and ordinal type columns from the original dataset.
  • Binary categorical features (all numerically-encoded).
  • Engineered features from other multi-level categorical features and mixed features.

Make sure that for any new columns that you have engineered, that you've excluded the original columns from the final dataset. Otherwise, their values will interfere with the analysis later on the project. For example, you should not keep "PRAEGENDE_JUGENDJAHRE", since its values won't be useful for the algorithm: only the values derived from it in the engineered features you created should be retained. As a reminder, your data should only be from the subset with few or no missing values.

In [2133]:
# If there are other re-engineering tasks you need to perform, make sure you
# take care of them here. (Dealing with missing data will come in step 2.1.)

pop.keys()
Out[2133]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'FINANZ_MINIMALIST',
       'FINANZ_SPARER', 'FINANZ_VORSORGER', 'FINANZ_ANLEGER',
       'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER', 'HEALTH_TYP',
       'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM',
       'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL',
       'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM',
       'SEMIO_PFLICHT', 'SEMIO_TRADV', 'VERS_TYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'KONSUMNAEHE', 'CAMEO_INTL_2015', 'KBA05_ANTG1',
       'KBA05_ANTG2', 'KBA05_ANTG3', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE',
       'INNENSTADT', 'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO',
       'ONLINE_AFFINITAET', 'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2',
       'PLZ8_ANTG3', 'PLZ8_ANTG4', 'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT',
       'ORTSGR_KLS9', 'RELAT_AB', 'wealth', 'life_stage', 'generation',
       'movement'],
      dtype='object')
In [2159]:
pop.shape
Out[2159]:
(891221, 85)
In [2137]:
# Do whatever you need to in order to ensure that the dataframe only contains
# the columns that should be passed to the algorithm functions.

Step 1.3: Create a Cleaning Function¶

Even though you've finished cleaning up the general population demographics data, it's important to look ahead to the future and realize that you'll need to perform the same cleaning steps on the customer demographics data. In this substep, complete the function below to execute the main feature selection, encoding, and re-engineering steps you performed above. Then, when it comes to looking at the customer data in Step 3, you can just run this function on that DataFrame to get the trimmed dataset in a single step.

In [2720]:
def clean_data(df:pd.DataFrame) -> pd.DataFrame:
    """
    Perform feature trimming, re-encoding, and engineering for population data
    
    INPUT: Sample DataFrame
    OUTPUT: Trimmed, cleaned and Median Imputed DataFrame
    """
    # Put in code here to execute all main cleaning steps:
    for x in range(len(feat_info)):
    # using regex to extract the digits
        missing_data = re.sub('[\[|\]]', '',feat_info.iloc[x]['missing_or_unknown']).split(',')
        # return the digit values for integer strings only
        if missing_data != ['']:
            missing_data = [np.int64(data) if (data!='X' and data!='XX') else data for data in missing_data]
            df = df.replace({feat_info.iloc[x]['attribute']: missing_data}, np.nan)

    df['CAMEO_INTL_2015'] = pd.to_numeric(df['CAMEO_INTL_2015'], downcast='integer', errors='coerce')    
    df['CAMEO_DEUG_2015'] = pd.to_numeric(df['CAMEO_DEUG_2015'], downcast='integer', errors='coerce')
    
    df = df.drop(columns=['TITEL_KZ', 'CAMEO_DEU_2015','AGER_TYP','KK_KUNDENTYP','KBA05_BAUMAX','GEBURTSJAHR','ALTER_HH', 'KBA13_ANZAHL_PKW', 'KBA05_ANTG3'])
    
    df['OST_WEST_KZ'] = df['OST_WEST_KZ'].replace({'W':0, 'O':1})
    df['OST_WEST_KZ'] = pd.to_numeric(df['OST_WEST_KZ'], errors='coerce')   #setting the errors to Nans
    
    df_outliers = identify_outlier_columns(df)
    df = df.drop(columns=df_outliers)
    # select, re-encode, and engineer column values.
    multi_col = []
    for col in cat_dimensions:
        column_data = df.get(col)
        if column_data is not None:
            counte = column_data.value_counts().reset_index().iloc[:, 0].values
            #print(f'{col} has: {counte} values')
            if len(counte) > 2:
                multi_col.append(col)
    valid_columns = [col for col in multi_col if col in df.columns] 
    valid_columns.remove('CAMEO_INTL_2015')
    valid_columns.remove('PRAEGENDE_JUGENDJAHRE')
    df = df.drop(columns=valid_columns)
    
    wealth = []
    life_stage = []
    for i in df['CAMEO_INTL_2015'].astype(float):
        if i//10==1.0:
            wealth.append(0)    #wealthy households
        elif i//10==2.0:
            wealth.append(1)    #prosperous households
        elif i//10==3.0:
            wealth.append(2)    #comfortable households
        elif i//10==4.0:
            wealth.append(3)    #less affluent households
        else:
            wealth.append(4)    #poorer households
        if i%10==1.0:
            life_stage.append(0)      #pre-family couple
        elif i%10==2.0:
            life_stage.append(1)      #young couple
        elif i%10==3.0:
            life_stage.append(2)      #family with school age
        elif i%10==4.0:
            life_stage.append(3)      #mature couples
        else:
            life_stage.append(4)      #elder in retirement
    df['wealth'] = wealth      #encoding wealth with floor division
    df['life_stage'] = life_stage      #encoding wealth with reminder

    
    generation = []
    movement = []
    for i in df['PRAEGENDE_JUGENDJAHRE'].astype(float):
        if i == 1 or i == 2:
            generation.append(0)         #40s 
        elif i == 3 or i == 4:
            generation.append(1)         #50s 
        elif i == 5 or i == 6 or i == 7:
            generation.append(2)         #60s 
        elif i == 8 or i == 9:
            generation.append(3)         #90s 
        elif i == 14 or i == 15:
            generation.append(4)         #80s 
        else:
            generation.append(5)
    
        if i == 1 or i == 3 or i == 5 or i == 8 or i == 10 or i == 12 or i == 14:
            movement.append(0)      #mainstream movement
        else:
            movement.append(1)       #avantgarde movement 
    df['generation']=generation      #encoding generation
    df['movement']=movement      #encoding movement type

    mixed_to_drop = []
    for col in mixed_cols:
        if (col != 'PRAEGENDE_JUGENDJAHRE') and (col != 'CAMEO_INTL_2015'):
            mixed_to_drop.append(col)
    for col in mixed_to_drop:
        try:
            df = df.drop(columns=col)
        except KeyError:
            continue

    df_columns = df.keys().to_list()
    df = pd.DataFrame(df, columns=df_columns)
    return df # Returning our cleaned dataframe.
In [2722]:
#reinputing just to ensure we dont use dilluted data

pop = pd.read_csv('Udacity_AZDIAS_Subset.csv', sep=';')
In [2724]:
df = clean_data(pop)
In [2726]:
df.head()
Out[2726]:
ALTERSKATEGORIE_GROB ANREDE_KZ FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER HEALTH_TYP PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV VERS_TYP HH_EINKOMMEN_SCORE W_KEIT_KIND_HH KONSUMNAEHE CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB wealth life_stage generation movement
0 2.0 1 3 4 3 5 5 3 NaN NaN 5.0 2 6 7 5 1 5 3 3 4 7 6 6 5 3 NaN 2.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 1.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 4 4 5 1
1 1.0 2 1 5 2 5 4 5 3.0 14.0 1.0 5 4 4 3 1 2 2 3 6 4 7 4 7 6 2.0 6.0 3.0 1.0 51.0 0.0 0.0 1.0 6.0 3.0 8.0 3.0 2.0 1.0 3.0 3.0 2.0 3.0 2.0 1.0 5.0 4.0 3.0 5.0 4.0 4 0 4 0
2 3.0 2 1 4 1 2 3 5 3.0 15.0 3.0 4 1 3 3 4 4 6 3 4 7 7 7 3 3 1.0 4.0 3.0 5.0 24.0 1.0 3.0 3.0 2.0 4.0 4.0 4.0 2.0 3.0 2.0 2.0 3.0 3.0 1.0 0.0 4.0 4.0 3.0 5.0 2.0 1 3 4 1
3 4.0 2 4 2 5 2 1 2 2.0 8.0 2.0 5 1 2 1 4 4 7 4 3 4 4 5 4 4 1.0 1.0 NaN 4.0 12.0 4.0 1.0 4.0 4.0 2.0 6.0 4.0 NaN 4.0 1.0 NaN 2.0 2.0 2.0 0.0 3.0 4.0 2.0 3.0 3.0 0 1 3 0
4 3.0 1 4 3 4 1 3 2 3.0 8.0 5.0 6 4 4 2 7 4 4 6 2 3 2 2 4 2 2.0 5.0 2.0 4.0 43.0 1.0 4.0 3.0 2.0 5.0 1.0 5.0 3.0 3.0 5.0 5.0 2.0 4.0 2.0 1.0 3.0 3.0 4.0 6.0 5.0 3 2 3 0
In [2728]:
df.shape
Out[2728]:
(891221, 54)
In [2730]:
df.isna().sum().sum()     #still need to sort nans
Out[2730]:
2897715

Confirming how cleaning function works just like we intend to¶

In [2733]:
assert pop.shape[0] == df.shape[0], "clean_data function is not working properly, rows mismatch"
print("If this is all you see, you passed the test")

If this is all you see, you passed the test

Step 2: Feature Transformation¶

Step 2.1: Apply Feature Scaling¶

Before we apply dimensionality reduction techniques to the data, we need to perform feature scaling so that the principal component vectors are not influenced by the natural differences in scale for features. Starting from this part of the project, you'll want to keep an eye on the API reference page for sklearn to help you navigate to all of the classes and functions that you'll need. In this substep, you'll need to check the following:

  • sklearn requires that data not have missing values in order for its estimators to work properly. So, before applying the scaler to your data, make sure that you've cleaned the DataFrame of the remaining missing values. This can be as simple as just removing all data points with missing data, or applying an SimpleImputer to replace all missing values. You might also try a more complicated procedure where you temporarily remove missing values in order to compute the scaling parameters before re-introducing those missing values and applying imputation. Think about how much missing data you have and what possible effects each approach might have on your analysis, and justify your decision in the discussion section below.
  • For the actual scaling function, a StandardScaler instance is suggested, scaling each feature to mean 0 and standard deviation 1.
  • For these classes, you can make use of the .fit_transform() method to both fit a procedure to the data as well as apply the transformation to the data at the same time. Don't forget to keep the fit sklearn objects handy, since you'll be applying them to the customer demographics data towards the end of the project.
In [2736]:
imp_median = SimpleImputer(strategy='median')         #median is less sensitive to outliers compared to mean
In [2738]:
numer_cols = df.select_dtypes(include='number').keys()
numer_cols
Out[2738]:
Index(['ALTERSKATEGORIE_GROB', 'ANREDE_KZ', 'FINANZ_MINIMALIST',
       'FINANZ_SPARER', 'FINANZ_VORSORGER', 'FINANZ_ANLEGER',
       'FINANZ_UNAUFFAELLIGER', 'FINANZ_HAUSBAUER', 'HEALTH_TYP',
       'PRAEGENDE_JUGENDJAHRE', 'RETOURTYP_BK_S', 'SEMIO_SOZ', 'SEMIO_FAM',
       'SEMIO_REL', 'SEMIO_MAT', 'SEMIO_VERT', 'SEMIO_LUST', 'SEMIO_ERL',
       'SEMIO_KULT', 'SEMIO_RAT', 'SEMIO_KRIT', 'SEMIO_DOM', 'SEMIO_KAEM',
       'SEMIO_PFLICHT', 'SEMIO_TRADV', 'VERS_TYP', 'HH_EINKOMMEN_SCORE',
       'W_KEIT_KIND_HH', 'KONSUMNAEHE', 'CAMEO_INTL_2015', 'KBA05_ANTG1',
       'KBA05_ANTG2', 'KBA05_GBZ', 'BALLRAUM', 'EWDICHTE', 'INNENSTADT',
       'GEBAEUDETYP_RASTER', 'KKK', 'MOBI_REGIO', 'ONLINE_AFFINITAET',
       'REGIOTYP', 'PLZ8_ANTG1', 'PLZ8_ANTG2', 'PLZ8_ANTG3', 'PLZ8_ANTG4',
       'PLZ8_HHZ', 'PLZ8_GBZ', 'ARBEIT', 'ORTSGR_KLS9', 'RELAT_AB', 'wealth',
       'life_stage', 'generation', 'movement'],
      dtype='object')
In [2740]:
df[numer_cols] = imp_median.fit_transform(df[numer_cols])
In [2742]:
df.isna().sum().sum()              #verifying simple imputer
Out[2742]:
0

Sample Scaling

In [2745]:
scaler = StandardScaler()
In [2747]:
# Apply feature scaling to the general population data.
scaled_data = scaler.fit_transform(df)
In [2749]:
scaled_data
Out[2749]:
array([[-0.75167967, -1.045218  , -0.05641562, ...,  1.21814982,
         1.10233491,  1.46453679],
       [-1.74333933,  0.95673821, -1.57035806, ..., -1.36014202,
         0.39299478, -0.68280975],
       [ 0.23997999,  0.95673821, -1.57035806, ...,  0.57357686,
         0.39299478,  1.46453679],
       ...,
       [-0.75167967,  0.95673821, -0.81338684, ...,  0.57357686,
         0.39299478, -0.68280975],
       [-1.74333933, -1.045218  , -1.57035806, ..., -1.36014202,
         0.39299478, -0.68280975],
       [ 1.23163965, -1.045218  ,  0.7005556 , ..., -0.0709961 ,
        -1.7350256 , -0.68280975]])
In [2751]:
type(scaled_data)
Out[2751]:
numpy.ndarray

Discussion 2.1: Apply Feature Scaling¶

Scaled data is a list as expected. Now we can move on

Step 2.2: Perform Dimensionality Reduction¶

On your scaled data, you are now ready to apply dimensionality reduction techniques.

  • Use sklearn's PCA class to apply principal component analysis on the data, thus finding the vectors of maximal variance in the data. To start, you should not set any parameters (so all components are computed) or set a number of components that is at least half the number of features (so there's enough features to see the general trend in variability).
  • Check out the ratio of variance explained by each principal component as well as the cumulative variance explained. Try plotting the cumulative or sequential values using matplotlib's plot() function. Based on what you find, select a value for the number of transformed features you'll retain for the clustering part of the project.
  • Once you've made a choice for the number of components to keep, make sure you re-fit a PCA instance to perform the decided-on transformation.
In [2755]:
# We aim to apply PCA to the data, but first, define a funcrtion to do all of these for us
In [2757]:
df.isna().sum().sum()
Out[2757]:
0
In [2759]:
def apply_pca(x:int=None, df:pd.DataFrame=df):
    """
    INPUT:
    x: str denoting number of components to be broken down into
    df: scaled dataframe to find principal components
    OUTPUT:
    df_pca: (np.array) denoting result
    pca: object denoting PCA's instantiation
    """
    scaler = StandardScaler()    #including scaling in function for ease
    scaled_data = scaler.fit_transform(df)
    pca = PCA(n_components=x)
    df_pca = pca.fit_transform(scaled_data)
    return pca, df_pca
In [2761]:
pca, df_pca = apply_pca()         #base pca with 0 components
In [2763]:
pca
Out[2763]:
PCA()
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
PCA()
In [2765]:
def scree_plot(pca):
    '''
    Creates a scree plot associated with the principal components 
    
    INPUT: pca - the result of instantian of PCA in scikit learn
            
    OUTPUT:
            None
    '''
    num_components=len(pca.explained_variance_ratio_)
    ind = np.arange(num_components)
    vals = pca.explained_variance_ratio_
 
    plt.figure(figsize=(15, 11))
    ax = plt.subplot(111)
    cumvals = np.cumsum(vals)
    ax.bar(ind, vals)
    ax.plot(ind, cumvals)
    for i in range(num_components):
        ax.annotate(r"%s%%" % ((str(vals[i]*100)[:4])), (ind[i]+0.2, vals[i]), va="bottom", ha="center", fontsize=12)
 
    ax.xaxis.set_tick_params(width=0)
    ax.yaxis.set_tick_params(width=2, length=12)
 
    ax.set_xlabel("Principal Component")
    ax.set_ylabel("Variance-Explained Ratio")
    plt.title('Explained Variance Per Principal Component')
In [2767]:
scree_plot(pca)
No description has been provided for this image
In [2769]:
def pca_results(full_dataset:pd.DataFrame, pca):
	'''
	Create a DataFrame of the PCA results
	Includes dimension feature weights and explained variance
	Visualizes the PCA results
	'''

	# Dimension indexing
	dimensions = ['Dimension {}'.format(i) for i in range(1,len(pca.components_)+1)]

	# PCA components as a datafra e
	components = pd.DataFrame(np.round(pca.components_, 4), columns = full_dataset.keys())
	components.index = dimensions

	# PCA explained variance as another dataframe
	ratios = pca.explained_variance_ratio_.reshape(len(pca.components_), 1)
	variance_ratios = pd.DataFrame(np.round(ratios, 4), columns = ['Explained Variance'])
	variance_ratios.index = dimensions

	# Create a bar plot visualization
	fig, ax = plt.subplots(figsize = (15,10))

	# Plot the feature weights as a function of the components
	components.plot(ax = ax, kind = 'bar');
	ax.set_ylabel("Feature Weights")
	ax.set_xticklabels(dimensions, rotation=0)


	# Display the explained variance ratios
	for i, ev in enumerate(pca.explained_variance_ratio_):
		ax.text(i-0.40, ax.get_ylim()[1] + 0.05, "Explained Variance\n          %.4f"%(ev))

	# Return a concatenated DataFrame
	return pd.concat([variance_ratios, components], axis = 1)
In [2771]:
def pca_computer(x:int=None, df:pd.DataFrame=df):
    """
    COmbination of the apply_pca and pca_result functions
    INPUT: 
    x(str): Number of principal components to split df into
    df(pd.DataFrame): Dataframe do compute pca on

    OUTPUT: 
    zing: Nonetype denoting pca component analysis
    """
    pca, df_pca = apply_pca()
    zing = pca_results(df, pca)
    return zing
In [2773]:
#loop to investigate variability threshold
for comp in range(3, df.shape[1]):
    pca, X_pca = apply_pca(comp, df)
    comp_check = pca_results(df, pca)
    if comp_check['Explained Variance'].sum() > 0.95:   # Re-apply PCA to the data while selecting for number of components to retain.
        break      
        

num_comps = comp_check.shape[0]
print("Using {} components, we can explain {}% of the variability in the original data.".format(comp_check.shape[0],comp_check['Explained Variance'].sum()))

/var/folders/y1/2cln1s193dgc6xy3yzry761r0000gn/T/ipykernel_845/3037338320.py:21: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`). Consider using `matplotlib.pyplot.close()`.
  fig, ax = plt.subplots(figsize = (15,10))

Using 36 components, we can explain 0.9503% of the variability in the original data.
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image
In [2774]:
#Manually confirming this:
In [2777]:
pca, df_pca = apply_pca(36, df)
In [2778]:
pca
Out[2778]:
PCA(n_components=36)
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
PCA(n_components=36)
In [2781]:
pca_results(df, pca)  # Investigating the variance accounted for by each principal component.
Out[2781]:
Explained Variance ALTERSKATEGORIE_GROB ANREDE_KZ FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER HEALTH_TYP PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV VERS_TYP HH_EINKOMMEN_SCORE W_KEIT_KIND_HH KONSUMNAEHE CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB wealth life_stage generation movement
Dimension 1 0.1868 0.1657 -0.0068 0.2295 -0.1946 0.1604 -0.1163 -0.1276 -0.1310 -0.0368 -0.1453 0.0501 -0.0333 -0.1061 -0.1553 -0.1093 0.0563 0.0989 0.1146 -0.0891 -0.1315 -0.0124 -0.0359 -0.0465 -0.1599 -0.1235 -0.0394 -0.1494 -0.0046 0.1607 -0.1982 0.2083 -0.0174 0.1961 0.1160 -0.1890 0.1542 0.1050 -0.0387 0.2211 0.0126 -0.0665 0.2175 -0.1559 -0.2205 -0.2102 -0.0530 0.1544 -0.1332 -0.1877 -0.1297 -0.1999 0.0952 -0.1353 0.0438
Dimension 2 0.1565 -0.2288 -0.0522 -0.0296 0.2119 -0.2160 0.2056 0.2107 -0.1489 0.0504 0.2384 -0.1630 0.0471 0.1445 0.2130 0.1391 0.0168 -0.1590 -0.1931 0.1615 0.1631 -0.0604 -0.0321 -0.0690 0.2043 0.2029 -0.0221 -0.1113 -0.1339 0.1107 -0.1248 0.1196 -0.0065 0.1296 0.0970 -0.1397 0.1179 0.0747 -0.0061 0.1417 0.1650 -0.0310 0.1412 -0.1037 -0.1415 -0.1319 -0.0207 0.1008 -0.1059 -0.1404 -0.1015 -0.1039 0.0483 0.2014 0.0452
Dimension 3 0.1102 -0.0264 0.3586 -0.1161 0.0702 -0.0532 0.1660 0.0923 0.0422 0.0019 0.0434 -0.0721 -0.2941 -0.2405 -0.0665 -0.0767 -0.3534 -0.0032 0.1879 -0.2920 0.1667 0.3105 0.3339 0.3603 0.0344 0.0073 0.0428 -0.0469 -0.0702 0.0301 -0.0184 0.0155 -0.0053 0.0207 0.0378 -0.0307 0.0395 0.0364 0.0110 0.0318 0.0160 0.0052 0.0319 -0.0184 -0.0239 -0.0236 0.0109 0.0214 -0.0345 -0.0375 -0.0271 0.0056 0.0654 0.0843 0.0480
Dimension 4 0.0505 -0.0013 -0.0399 0.1386 -0.0205 0.0093 -0.1181 0.1021 -0.1463 -0.0104 0.0228 0.0578 -0.0115 0.0564 0.0470 0.0410 -0.0143 0.0710 -0.0718 -0.0427 -0.0221 0.0370 0.0830 0.0437 -0.0150 -0.0086 0.0777 -0.3376 -0.0598 -0.1469 -0.1833 0.1123 -0.0345 0.0768 -0.2395 0.2645 -0.2268 -0.0830 -0.3258 0.0988 0.0963 -0.2598 -0.0439 0.0913 0.0987 0.0881 0.1620 0.0360 0.0303 0.2396 0.0674 -0.1461 0.1560 0.0721 0.3908
Dimension 5 0.0412 0.0809 -0.1529 0.1052 0.0498 0.0469 0.1670 0.1747 -0.0482 -0.0912 -0.1143 0.0827 -0.1076 0.2167 0.2000 0.0478 -0.1048 0.1984 -0.1269 -0.0363 -0.1865 0.1314 -0.0556 -0.0233 -0.0253 -0.2390 0.2252 -0.2022 0.0998 0.0440 0.1648 -0.0609 -0.0501 -0.0239 0.1991 -0.0847 0.1460 0.1178 0.1912 0.0031 -0.1688 0.1846 -0.0551 0.0587 0.0761 0.0644 0.0655 -0.0501 -0.0484 -0.1188 -0.0269 0.2616 0.2470 0.1202 0.2837
Dimension 6 0.0343 0.0130 -0.0313 0.1326 0.0031 -0.0453 0.0270 -0.0206 -0.2173 0.1514 0.0779 -0.0151 0.0335 -0.0632 -0.0456 -0.1917 0.0283 -0.0687 0.0470 0.0583 -0.0693 0.0017 0.0137 0.0595 0.0011 -0.0916 0.0939 -0.0490 -0.1577 0.1331 -0.0249 0.1532 -0.1381 0.1059 -0.1502 0.1515 -0.1692 0.1518 0.4030 0.1355 0.1604 0.3764 0.0262 -0.0825 -0.0268 -0.0014 -0.2370 -0.1901 0.2724 0.2077 0.2296 -0.0277 -0.0116 0.0674 0.0006
Dimension 7 0.0281 -0.1349 0.0892 0.0929 -0.1620 0.1609 -0.1080 -0.0949 0.0868 -0.2773 -0.2354 0.1153 -0.0950 0.0690 0.2015 0.3237 -0.0774 -0.0125 -0.0760 0.0896 0.2884 0.0253 0.0369 0.0872 0.2106 0.2407 -0.2502 -0.0436 0.2097 0.0885 0.0011 0.0869 0.0603 0.1104 -0.0232 0.0616 -0.0453 0.1612 0.0831 0.1064 -0.1696 0.1051 0.0403 0.0157 -0.0302 -0.0731 -0.0940 -0.0261 0.1894 0.0680 0.1637 0.0089 0.1287 -0.1816 0.0213
Dimension 8 0.0270 -0.0928 -0.0245 -0.0226 -0.0139 0.0870 -0.0392 -0.0128 0.0143 0.2703 -0.0409 0.0926 0.0301 0.0046 0.0312 -0.1183 -0.0290 -0.0778 -0.0366 0.0750 0.0148 0.0476 0.0455 0.0198 0.0900 0.0149 0.2431 0.0581 0.1289 -0.0760 -0.0370 0.0448 0.2253 0.0324 -0.0683 0.0950 -0.0501 0.0730 0.1849 0.0233 -0.1059 0.2290 0.1241 0.0955 -0.0310 -0.1211 0.5253 0.5097 -0.0092 0.0662 0.0018 -0.0473 -0.0685 -0.0477 -0.0699
Dimension 9 0.0251 -0.1564 -0.0239 -0.0391 -0.0318 0.0902 -0.0259 -0.0627 0.1277 0.2989 -0.1028 0.1635 0.0495 0.0598 0.1142 -0.1082 -0.1013 -0.0789 -0.1171 0.0457 0.0482 0.0945 0.0594 0.0068 0.1140 0.0644 0.3415 -0.0389 0.2427 0.0235 -0.1127 -0.0160 -0.2990 -0.1070 -0.1475 0.0147 -0.1086 -0.1469 -0.0358 -0.0724 -0.2071 -0.0219 0.0904 -0.2772 -0.1257 0.0317 -0.2478 -0.2032 -0.2015 0.0073 -0.2429 -0.1057 0.0523 -0.0795 -0.0267
Dimension 10 0.0242 -0.0721 -0.0439 0.0051 -0.0006 0.0037 -0.0245 -0.0518 0.0405 0.3397 -0.0012 0.0024 0.0238 0.0184 -0.0131 -0.0822 -0.0401 -0.0995 -0.0719 -0.0079 -0.0395 0.0418 0.1013 -0.0531 -0.0059 0.0129 0.3056 0.0569 0.0573 0.0653 0.0959 0.0272 0.3496 0.1612 0.1920 -0.0612 0.1423 0.1156 -0.2434 0.0904 -0.1018 -0.3077 -0.0344 0.1701 -0.0004 -0.1273 -0.2598 -0.0998 0.2817 -0.0483 0.3314 0.0796 0.0075 -0.0198 0.0147
Dimension 11 0.0227 0.0501 0.0178 -0.3120 0.1885 -0.1162 0.1338 0.1671 0.3693 -0.2981 0.0973 -0.2022 0.0625 -0.0134 0.0196 -0.1576 -0.0074 0.2382 0.0208 -0.0397 -0.1306 -0.1006 -0.1486 -0.0586 -0.0496 -0.1180 0.0793 0.0551 0.2985 0.0660 -0.1242 0.0408 0.0965 0.0154 -0.1357 0.0696 -0.0984 0.0555 -0.0267 0.0114 -0.0622 0.0010 0.1620 -0.1303 -0.1509 -0.1547 -0.0828 0.0711 0.1741 0.1131 0.1098 -0.1457 0.0843 -0.0657 0.1111
Dimension 12 0.0199 0.0448 -0.0040 0.0113 0.0268 -0.0111 -0.0101 0.0683 -0.0608 -0.0213 0.0164 0.0543 -0.0462 0.0044 0.0141 0.0651 -0.0061 0.1476 0.0406 0.0188 0.0062 0.0331 -0.0169 0.0025 0.0364 -0.0307 0.0125 0.0023 0.0547 -0.1729 -0.0424 0.0365 0.5722 0.1873 -0.1715 0.0503 -0.1199 -0.1604 0.0977 0.1154 -0.0043 0.1985 -0.0339 0.2503 0.0123 -0.1031 -0.3063 -0.1864 -0.3271 -0.0189 -0.2877 -0.0342 -0.1448 0.0480 0.0108
Dimension 13 0.0179 0.0220 -0.0455 -0.0493 -0.0023 0.0464 -0.0299 0.0157 -0.0041 0.0114 -0.0101 0.1670 -0.0278 0.0210 -0.0226 0.0612 0.0415 -0.1253 -0.0231 -0.0228 -0.0372 0.0083 0.0039 0.0151 -0.0561 -0.0304 -0.0160 0.0459 -0.2407 0.3070 -0.1223 -0.3035 0.2646 -0.3014 -0.1625 0.0741 -0.1095 0.5003 -0.0124 -0.2345 0.0356 -0.0751 0.0575 0.0354 -0.0077 -0.0636 -0.0610 -0.0163 -0.0874 0.0376 -0.1306 -0.1197 0.3114 0.0596 -0.0403
Dimension 14 0.0173 0.1426 -0.0451 0.0098 0.0684 -0.1130 -0.0115 0.0210 -0.1741 0.0360 0.0752 0.3736 -0.0762 0.0500 -0.0005 0.2211 0.0118 0.1608 0.1022 0.0041 -0.0017 0.0474 0.0343 0.0150 -0.0033 -0.0417 -0.0901 0.1194 -0.1335 0.0092 0.1732 -0.0479 0.0053 -0.0417 -0.0971 0.0418 -0.0978 -0.1235 -0.0891 -0.0694 -0.2784 -0.0748 0.2400 -0.2787 -0.2757 -0.2033 -0.0599 0.1358 0.1060 0.0454 0.1036 0.1741 -0.2588 0.2701 0.0928
Dimension 15 0.0149 -0.0771 0.0276 0.0035 -0.0758 0.2070 -0.0693 -0.0942 0.1716 -0.1809 -0.1029 -0.1433 0.0014 0.0307 0.1042 -0.0752 -0.0806 0.0803 -0.0688 -0.0061 0.0330 -0.0963 -0.0065 -0.0005 0.0905 0.0264 0.2847 -0.0069 -0.2821 0.0597 0.3561 0.0565 -0.0056 0.0353 -0.1441 0.0480 -0.0771 0.1422 -0.0647 0.0283 0.3863 -0.0986 0.1096 -0.0853 -0.1041 -0.0881 -0.0198 0.0582 -0.0998 0.0423 -0.1413 0.3516 -0.2286 -0.1271 0.0959
Dimension 16 0.0143 0.0138 0.0221 -0.0171 0.0819 -0.1219 0.0063 0.0872 -0.0916 0.0567 0.0936 -0.0021 0.0703 -0.0568 -0.0769 -0.0495 0.0219 -0.0616 0.0210 -0.0225 -0.0239 0.0069 0.0072 -0.0088 -0.0550 -0.0060 -0.1558 -0.0116 0.3251 0.2163 0.1419 0.1778 -0.1983 0.1367 -0.1194 0.0408 -0.1283 0.5094 -0.1243 0.1339 -0.2410 -0.1004 -0.1039 0.1448 0.1184 0.0776 0.0442 -0.0232 -0.1381 0.0079 -0.2045 0.1427 -0.3283 0.0129 0.0155
Dimension 17 0.0132 -0.0513 0.0724 -0.1327 0.0444 0.0227 0.0857 -0.0353 0.1319 -0.2714 0.0213 0.5456 0.0844 -0.0428 0.0738 -0.1402 -0.0260 0.0862 0.0723 0.0274 0.0690 -0.1567 -0.0839 0.0109 0.1151 0.0292 0.2423 0.0742 -0.2950 0.0297 -0.2095 0.1166 -0.1142 0.0546 0.1951 -0.0073 0.0356 0.0845 0.0432 0.0630 -0.0818 -0.0341 -0.1135 0.1783 0.1249 0.0853 0.0100 -0.0381 -0.0312 -0.0878 0.0692 -0.2188 -0.2321 0.0457 0.0569
Dimension 18 0.0128 0.1220 -0.0376 0.0012 -0.0203 0.0020 -0.0649 -0.0815 0.0760 0.2913 -0.0188 -0.1724 0.0093 0.1033 0.0271 0.3524 -0.0068 0.5576 0.1496 0.0511 0.0757 0.0708 0.0926 0.0015 0.0487 0.0846 0.1185 -0.0544 0.0831 0.1586 -0.1931 -0.0804 -0.0482 -0.0989 0.0719 -0.0479 0.0612 0.1565 -0.0182 -0.0875 0.2291 0.0893 -0.0686 -0.0026 0.0566 0.0748 0.0243 -0.0185 0.0640 -0.0139 0.0302 -0.1875 -0.2939 0.0288 0.0158
Dimension 19 0.0115 0.0630 -0.0366 -0.0707 0.0647 -0.1692 0.0992 -0.0731 0.0998 0.1374 -0.0165 0.1091 0.0145 0.0321 -0.0165 0.1862 0.0039 0.2871 0.0808 0.0297 0.0003 -0.0186 0.0176 -0.0211 -0.0192 0.0113 0.0781 0.2094 -0.1864 -0.0126 0.1303 0.1879 -0.0923 0.2802 -0.1437 -0.0245 -0.1561 -0.0046 -0.0227 0.2086 -0.0169 -0.0915 -0.1194 -0.0176 0.0617 0.0495 0.0969 0.0192 0.1029 -0.0319 -0.1996 0.0798 0.4844 -0.0583 -0.3522
Dimension 20 0.0106 0.0316 -0.0486 0.1303 0.0091 -0.1137 -0.1323 0.1284 -0.3322 -0.4147 0.0159 -0.1673 -0.1159 0.0035 -0.0557 0.1537 0.0748 -0.0468 0.0366 -0.0450 0.0527 0.1790 -0.0644 -0.0117 -0.1139 0.1457 0.5280 0.0596 0.0727 -0.0173 -0.0549 -0.0397 -0.1086 -0.0479 -0.0749 0.0361 0.0157 0.0163 -0.0359 -0.0396 -0.0982 -0.0320 0.0137 0.0989 -0.0264 -0.0637 -0.0236 0.0224 -0.0372 -0.0015 0.1014 -0.0432 -0.0436 -0.0301 -0.3711
Dimension 21 0.0101 -0.0132 0.0482 -0.0466 -0.0118 0.0490 -0.1228 0.0997 -0.0609 0.0917 0.0449 0.0158 0.2853 -0.0643 -0.0316 -0.0425 0.0430 0.1893 0.0921 0.0287 0.0554 -0.0410 0.0661 0.0661 0.1033 0.0405 -0.0681 0.1150 0.0446 -0.3308 0.1049 0.0910 -0.3062 -0.1780 0.0409 0.0828 0.1115 0.1443 -0.0452 -0.0986 0.0657 0.0781 0.1865 0.4081 -0.0761 -0.3004 -0.2055 0.0707 -0.1655 -0.0308 0.1056 0.0493 0.2475 0.0496 0.0639
Dimension 22 0.0092 -0.0180 -0.0615 -0.0807 0.0991 -0.0485 0.0611 0.2360 0.0468 0.0888 -0.0440 0.4282 -0.2997 0.0224 -0.0099 0.0589 -0.0210 -0.0650 -0.1413 -0.0461 -0.0583 0.1015 -0.0815 -0.0490 -0.1547 -0.0643 -0.1021 -0.1202 0.2391 -0.0646 0.0054 -0.0288 -0.0937 -0.0023 0.0958 0.0381 -0.1465 0.0079 -0.0751 0.0188 0.5184 0.0259 0.0380 0.0601 -0.0451 -0.1003 -0.0408 0.0258 0.0038 -0.0855 0.1162 0.0345 -0.0807 -0.2151 -0.2306
Dimension 23 0.0091 -0.0744 0.1077 0.1088 -0.0379 -0.1414 -0.0385 -0.0832 -0.1755 -0.1512 0.1331 0.2489 0.4365 -0.1499 -0.0945 -0.0876 0.0729 0.0803 0.1283 -0.0278 0.1095 0.0308 -0.0008 0.0922 0.0682 0.1359 0.1201 -0.0393 0.3651 0.1861 0.1546 -0.0884 0.2030 -0.0250 0.1513 -0.0207 -0.0685 -0.0916 -0.0275 -0.0459 0.2865 0.0260 -0.0486 -0.1863 0.0098 0.0975 0.0682 -0.0044 0.0788 -0.0553 -0.1005 0.0979 0.1712 0.0914 0.0856
Dimension 24 0.0087 -0.0177 0.0145 -0.0077 -0.0086 0.0428 0.0091 -0.0327 0.0004 0.0157 0.0067 0.0178 -0.0487 -0.0158 0.0251 -0.0455 -0.0298 0.0001 -0.0136 -0.0027 -0.0065 0.0047 -0.0410 -0.0417 0.0551 -0.0234 -0.0323 -0.0044 0.0265 0.6116 0.0304 -0.0568 -0.1622 -0.0231 -0.3029 -0.1338 0.0371 -0.4271 -0.0167 -0.0150 0.0279 0.0237 -0.0898 0.4421 0.0160 -0.2249 0.0004 0.1057 0.0744 -0.0536 -0.0047 0.0165 0.0161 0.0491 0.0583
Dimension 25 0.0080 0.0508 0.0217 -0.0784 0.0586 -0.0331 0.0309 0.2325 0.0851 -0.0172 -0.1972 0.0044 0.6089 0.1538 0.0269 0.1455 -0.0024 -0.1053 0.0204 -0.0794 -0.1076 0.4226 0.0663 0.1053 -0.0284 -0.1095 -0.0927 -0.1236 -0.2202 0.1140 -0.0364 -0.0135 0.0287 0.0814 -0.0290 0.0527 0.0537 -0.0829 0.0356 0.0498 -0.0401 -0.0236 0.0488 -0.0150 -0.0170 0.0166 0.0083 -0.0050 -0.0883 0.0188 0.1156 0.0151 -0.1184 -0.2595 -0.1109
Dimension 26 0.0077 0.1404 -0.0590 -0.1254 -0.0307 0.0115 -0.0318 0.0176 0.2612 0.0518 0.0873 -0.0159 -0.0111 -0.0132 -0.1890 0.3954 0.1616 -0.3369 0.1465 -0.0101 0.0825 -0.0793 0.0538 0.0174 -0.3369 0.1488 0.1986 0.0229 0.0079 0.0778 0.0445 0.0751 -0.0856 0.0535 0.2714 0.1108 -0.1262 -0.0964 0.1987 0.0378 0.0050 0.0262 0.0474 0.0908 -0.0093 -0.0751 -0.0395 0.0258 -0.0312 0.0907 -0.1444 0.0216 0.0448 -0.0079 0.3267
Dimension 27 0.0074 -0.0432 -0.0030 0.0174 0.0102 -0.0365 -0.1380 0.2001 -0.1075 0.0605 -0.0550 -0.1465 -0.1081 0.0389 0.0473 -0.1995 -0.0001 0.2128 -0.0266 0.0417 -0.0028 -0.1324 0.0316 0.0636 0.0651 0.0560 -0.0832 0.0373 -0.1159 0.3172 0.0340 0.0827 0.0608 -0.0910 0.5142 0.2545 -0.3735 -0.1572 -0.0606 -0.0145 -0.1497 -0.0006 0.0972 0.0271 -0.0342 0.0564 0.0259 -0.0293 -0.2008 0.0689 0.0623 0.0355 0.0487 -0.1398 -0.1206
Dimension 28 0.0070 0.0156 -0.0224 -0.0856 0.0583 -0.0929 -0.1230 0.3137 -0.0288 0.1099 -0.0302 0.1367 -0.1347 -0.0595 -0.0524 -0.0514 0.1107 0.0769 0.1395 0.0701 0.1267 -0.2394 0.0425 0.0212 -0.0966 0.1806 0.0331 -0.0205 -0.0112 0.1355 0.1006 0.0988 0.1045 -0.0412 -0.2757 0.0639 0.5154 -0.0376 0.0487 0.0281 -0.0167 0.0236 0.0487 -0.1431 0.0406 0.2203 0.0441 -0.0725 -0.2104 0.0745 0.1832 0.0720 0.0762 -0.2464 0.0882
Dimension 29 0.0066 0.0872 -0.0576 -0.1133 -0.0356 0.2669 -0.1574 -0.0655 0.0779 -0.0716 0.1146 0.0011 -0.0639 0.0296 0.0154 0.1020 0.0097 -0.1180 -0.0146 0.0409 -0.0971 0.2258 0.0023 -0.0422 0.2029 -0.2260 -0.0213 0.3560 0.1551 0.1576 -0.0247 0.1683 -0.0073 0.1384 0.0379 0.1799 0.1398 -0.0214 -0.0577 0.0786 0.1873 -0.0320 0.0506 -0.0777 -0.0031 0.1844 0.0549 -0.0552 -0.2977 0.0934 0.2058 -0.0277 0.0868 0.3736 -0.0674
Dimension 30 0.0059 0.0597 -0.1046 -0.0148 -0.0860 -0.0959 -0.1864 0.0279 0.1747 -0.0417 0.0615 0.0399 -0.1585 0.2114 0.0544 -0.3988 0.1818 0.0746 0.1197 0.1754 -0.0319 0.5452 -0.0647 -0.0716 -0.0663 0.4172 -0.1218 0.0360 -0.0708 -0.0394 0.0282 0.0411 -0.0035 -0.0011 0.0266 -0.0533 0.0358 0.0312 0.0914 0.0003 0.0276 -0.0752 -0.0269 0.0199 0.0141 -0.0327 -0.0136 0.0082 0.1019 0.0450 -0.1339 0.0052 0.0107 0.0782 0.0949
Dimension 31 0.0058 -0.0667 -0.0832 -0.0029 -0.0240 -0.0739 -0.0559 -0.0857 0.0477 -0.1671 0.0657 0.0936 0.0600 0.0059 -0.0118 -0.0136 -0.0549 0.1093 -0.4161 -0.0083 -0.1852 -0.0843 0.5210 0.0243 -0.2189 0.0400 -0.0509 -0.0233 0.0146 0.0550 -0.0316 0.1060 0.0098 -0.0598 0.0879 0.1492 0.3035 -0.0464 -0.0105 0.0024 0.0127 0.0773 0.0972 0.0562 0.0159 -0.0328 -0.0263 -0.0093 0.1568 0.2424 -0.2691 -0.0142 -0.0674 0.1063 -0.1859
Dimension 32 0.0056 0.1468 -0.0197 0.0452 0.1493 -0.2450 0.0739 -0.1011 -0.1370 -0.0377 -0.1541 0.0143 -0.1302 0.0947 0.0389 0.0269 0.0461 -0.1586 0.2798 0.1374 -0.0861 0.0026 0.0497 0.0239 0.4403 -0.2554 0.0267 0.0652 0.0886 -0.0133 -0.0152 0.0766 -0.0013 -0.1149 0.1171 0.1122 0.1763 -0.0227 0.0907 -0.0587 0.0999 -0.2332 0.0586 0.0849 0.0043 -0.0461 -0.0281 0.0055 0.1445 0.2420 -0.2782 -0.0329 -0.0245 -0.2634 0.0631
Dimension 33 0.0053 0.1259 0.0020 -0.0128 0.0345 0.0694 0.1570 -0.1202 0.0028 -0.0526 -0.0836 0.0029 0.0096 0.5478 -0.0467 -0.1397 -0.0095 -0.0831 0.2577 -0.0658 -0.1962 -0.3112 0.2348 0.2374 0.0060 0.2453 -0.0199 -0.1301 0.1381 -0.0129 -0.0155 -0.1691 -0.0330 0.1478 -0.0654 0.0490 -0.0483 0.0031 0.0751 0.0240 0.0545 -0.0934 -0.0297 0.0089 -0.0324 -0.0485 -0.0090 0.0299 -0.1481 -0.0731 0.1562 0.0146 0.0070 0.1625 -0.1535
Dimension 34 0.0051 0.0540 -0.0911 -0.0015 0.0689 0.1057 0.2760 -0.1692 -0.0498 -0.0475 -0.0330 -0.0198 0.0507 -0.0381 0.0886 -0.0066 -0.0350 0.0291 0.0429 0.1771 0.1416 0.0716 0.1457 -0.0777 -0.2167 -0.0196 -0.0218 -0.0885 0.0781 0.0027 0.0250 0.4515 0.1420 -0.4172 -0.1063 -0.1057 -0.1657 -0.0482 0.2283 -0.0127 0.0269 -0.2897 0.1777 0.0743 0.0492 0.1127 0.0367 -0.0534 -0.0588 -0.1536 0.1500 0.0168 -0.0111 0.0436 -0.0652
Dimension 35 0.0049 0.0249 -0.1455 0.0236 -0.0014 0.0277 -0.0449 0.0875 -0.0094 -0.0525 -0.0350 -0.0257 0.0470 -0.0787 0.2795 -0.1720 0.0099 0.0262 0.1347 0.3072 0.4620 -0.0248 0.2586 -0.0446 -0.2712 -0.3384 -0.0262 0.0400 0.0491 -0.0104 -0.0251 -0.3488 -0.0809 0.2530 0.0383 0.0135 -0.0111 0.0116 0.0808 0.0189 0.0322 -0.1452 -0.1002 -0.0253 -0.0159 -0.0914 -0.0181 0.0529 -0.0507 0.0747 -0.0030 -0.0176 -0.0032 0.0367 -0.0358
Dimension 36 0.0049 0.0839 -0.0754 -0.0258 0.0431 -0.0650 0.0364 0.0209 -0.0410 -0.0820 -0.0365 -0.0006 -0.0453 -0.2976 -0.1225 0.0456 0.1801 0.0713 -0.2387 0.0692 -0.2627 0.0474 0.4259 -0.1632 0.2694 0.1602 0.0115 0.1254 0.0062 -0.0279 -0.0025 -0.1814 -0.0542 0.1335 -0.0557 -0.1685 -0.1968 0.0183 0.2352 -0.0381 0.0074 -0.1351 -0.0998 -0.0507 0.0041 -0.0158 0.0026 0.0399 -0.1760 -0.1216 0.1900 -0.0257 0.0073 -0.1768 0.2009
No description has been provided for this image
In [2783]:
pca_results(df, pca)['Explained Variance'].sum()

#Confirmed
Out[2783]:
0.9503
No description has been provided for this image

Discussion 2.2: Perform Dimensionality Reduction¶

(Double-click this cell and replace this text with your own text, reporting your findings and decisions regarding dimensionality reduction. How many principal components / transformed features are you retaining for the next step of the analysis?)

• PCA being a component of data optimization using data reduction, in this particular evaluation, we are attempting to explain 95% of our data's variability.

• Using the For-Loop and a variance threshold, we were able to discover 36 n_components were just right.

Step 2.3: Interpret Principal Components¶

Now that we have our transformed principal components, it's a nice idea to check out the weight of each variable on the first few components to see if they can be interpreted in some fashion.

As a reminder, each principal component is a unit vector that points in the direction of highest variance (after accounting for the variance captured by earlier principal components). The further a weight is from zero, the more the principal component is in the direction of the corresponding feature. If two features have large weights of the same sign (both positive or both negative), then increases in one tend expect to be associated with increases in the other. To contrast, features with different signs can be expected to show a negative correlation: increases in one variable should result in a decrease in the other.

  • To investigate the features, you should map each weight to their corresponding feature name, then sort the features according to weight. The most interesting features for each principal component, then, will be those at the beginning and end of the sorted list. Use the data dictionary document to help you understand these most prominent features, their relationships, and what a positive or negative value on the principal component might indicate.
  • You should investigate and interpret feature associations from the first three principal components in this substep. To help facilitate this, you should write a function that you can call at any time to print the sorted list of feature weights, for the i-th principal component. This might come in handy in the next step of the project, when you interpret the tendencies of the discovered clusters.
In [2787]:
# HINT: Try defining a function here or in a new cell that you can reuse in the
# other cells.
In [2789]:
def weights(n:int, pca, df:pd.DataFrame) -> pd.DataFrame:
    """
    INPUT: 
    n(int): index ot principal component to analyze
    pca(sklearn.decomposition.PCA): a PCA fitted object containing eigen vectors
    df(pd.DataFrame): a dataframe to examine weights

    OUTPUT: 
    df(pd.DataFrame): result dataframe having columns as index, and corresponding weights
    """
    
    col_list = list(df.keys())
    weight=list(pca.components_[n])
    
    df =pd.DataFrame(list(zip(col_list, weight)))
    df = df.set_axis(['Dimension', 'Weights'], axis=1)
    df = df.sort_values(by=['Weights'] , ascending=False)
    df.set_index('Dimension', inplace=True)
    
    return df
In [2809]:
# Map weights for the first principal component to corresponding feature names
# and then print the linked values, sorted by weight.

first_comp = weights(0, pca, df)
In [2811]:
plt.figure(figsize=(10, len(first_comp) * 0.3))  # Adjust figure size based on number of dimensions
plt.bar(first_comp.index, first_comp['Weights'])
plt.title(f'Weights of 1 Principal Component')
plt.xlabel('Weight')
plt.ylabel('Dimension')
plt.xticks(rotation=90)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [2813]:
# Map weights for the second principal component to corresponding feature names
# and then print the linked values, sorted by weight.

second_comp = weights(1, pca, df)
In [2815]:
plt.figure(figsize=(10, len(second_comp) * 0.3))  # Adjust figure size based on number of dimensions
plt.bar(second_comp.index, second_comp['Weights'])
plt.title(f'Weights of 2 Principal Components')
plt.xlabel('Weight')
plt.ylabel('Dimension')
plt.xticks(rotation=90)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [2817]:
# Map weights for the third principal component to corresponding feature names
# and then print the linked values, sorted by weight.

third_component = weights(2, pca, df)
In [2819]:
plt.figure(figsize=(10, len(third_component) * 0.3))  # Adjust figure size based on number of dimensions
plt.bar(third_component.index, third_component['Weights'])
plt.title(f'Weights of 3 Principal Components')
plt.xlabel('Weight')
plt.ylabel('Dimension')
plt.xticks(rotation=90)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [2821]:
final_comp = weights(35, pca, df)       #being 0 index, our final component(37) == 36
In [2823]:
plt.figure(figsize=(10, len(final_comp) * 0.3))  # Adjust figure size based on number of dimensions
plt.bar(final_comp.index, final_comp['Weights'])
plt.title(f'Weights of 3 Principal Components')
plt.xlabel('Weight')
plt.ylabel('Dimension')
plt.xticks(rotation=90)
plt.tight_layout()
plt.show()
No description has been provided for this image
In [2839]:
final_comp
Out[2839]:
Weights
Dimension
SEMIO_DOM 0.425884
SEMIO_PFLICHT 0.269351
KKK 0.235192
movement 0.200948
RELAT_AB 0.190000
SEMIO_VERT 0.180109
SEMIO_TRADV 0.160221
KBA05_GBZ 0.133485
HH_EINKOMMEN_SCORE 0.125398
ALTERSKATEGORIE_GROB 0.083941
SEMIO_LUST 0.071323
SEMIO_KULT 0.069172
SEMIO_KRIT 0.047356
SEMIO_MAT 0.045609
FINANZ_SPARER 0.043139
PLZ8_GBZ 0.039872
FINANZ_ANLEGER 0.036410
FINANZ_UNAUFFAELLIGER 0.020877
GEBAEUDETYP_RASTER 0.018348
VERS_TYP 0.011514
ONLINE_AFFINITAET 0.007423
life_stage 0.007267
W_KEIT_KIND_HH 0.006190
PLZ8_ANTG3 0.004076
PLZ8_HHZ 0.002626
RETOURTYP_BK_S -0.000599
CAMEO_INTL_2015 -0.002522
PLZ8_ANTG4 -0.015840
wealth -0.025718
FINANZ_MINIMALIST -0.025776
KONSUMNAEHE -0.027935
PRAEGENDE_JUGENDJAHRE -0.036516
MOBI_REGIO -0.038148
FINANZ_HAUSBAUER -0.040972
SEMIO_SOZ -0.045268
PLZ8_ANTG2 -0.050656
KBA05_ANTG2 -0.054157
BALLRAUM -0.055670
FINANZ_VORSORGER -0.065035
ANREDE_KZ -0.075384
HEALTH_TYP -0.082011
PLZ8_ANTG1 -0.099785
ORTSGR_KLS9 -0.121589
SEMIO_REL -0.122525
REGIOTYP -0.135097
SEMIO_KAEM -0.163151
EWDICHTE -0.168451
ARBEIT -0.175954
generation -0.176797
KBA05_ANTG1 -0.181449
INNENSTADT -0.196824
SEMIO_ERL -0.238687
SEMIO_RAT -0.262656
SEMIO_FAM -0.297620

Discussion 2.3: Interpret Principal Components¶

  • First Component Weights: In this, FINANZ_MINIMALIST have the heightest weight, denoting Low finicial interest investors
  • Second Component Weights: For this also, PRAEGENDE_JUGENDJAHRE has the heighest values prioritising movement across generations
  • Third Component Weights:Here, SEMIO_KAEM tells about how Belligerent the customer is
  • Final Component Weights: Finally, the dominant minded(SEMIO_DOM), the dutiful(SEMIO_PFLICHT) and then people with very high purhcasing power(KKK)

These show there isnt any correlation between the different principal component weights

Step 3: Clustering¶

Step 3.1: Apply Clustering to General Population¶

You've assessed and cleaned the demographics data, then scaled and transformed them. Now, it's time to see how the data clusters in the principal components space. In this substep, you will apply k-means clustering to the dataset and use the average within-cluster distances from each point to their assigned cluster's centroid to decide on a number of clusters to keep.

  • Use sklearn's KMeans class to perform k-means clustering on the PCA-transformed data.
  • Then, compute the average difference from each point to its assigned cluster's center. Hint: The KMeans object's .score() method might be useful here, but note that in sklearn, scores tend to be defined so that larger is better. Try applying it to a small, toy dataset, or use an internet search to help your understanding.
  • Perform the above two steps for a number of different cluster counts. You can then see how the average distance decreases with an increasing number of clusters. However, each additional cluster provides a smaller net benefit. Use this fact to select a final number of clusters in which to group the data. Warning: because of the large size of the dataset, it can take a long time for the algorithm to resolve. The more clusters to fit, the longer the algorithm will take. You should test for cluster counts through at least 10 clusters to get the full picture, but you shouldn't need to test for a number of clusters above about 30.
  • Once you've selected a final number of clusters to use, re-fit a KMeans instance to perform the clustering operation. Make sure that you also obtain the cluster assignments for the general demographics data, since you'll be using them in the final Step 3.3.
In [2842]:
pca, input_pca = apply_pca(36, df)
In [2844]:
input_pca
Out[2844]:
array([[-1.4945402 ,  1.14600198,  1.7340266 , ...,  0.08530578,
         0.12782479, -0.04777715],
       [-4.43718914,  2.02626803,  1.95999335, ..., -0.16507349,
        -0.48791817,  0.48001781],
       [ 0.6756197 ,  0.51984735,  3.04428583, ..., -0.03045957,
         0.37642888,  0.42620528],
       ...,
       [ 0.86775767,  3.89047834,  1.84754791, ...,  0.59030547,
         0.17706409,  0.1290213 ],
       [-6.325277  ,  1.46714174, -3.27781371, ..., -0.41757441,
         0.03695333,  0.06304165],
       [ 2.02143688, -1.7472236 , -3.38823799, ...,  1.08966956,
         0.27419976,  0.07838945]])
In [2848]:
Elbow_M = KElbowVisualizer(KMeans(), k=36)     #using base 

# Fit the visualizer to the PCA-transformed data
Elbow_M.fit(input_pca)

# Display the Elbow Method plot
Elbow_M.show()
No description has been provided for this image
Out[2848]:
<Axes: title={'center': 'Distortion Score Elbow for KMeans Clustering'}, xlabel='k', ylabel='distortion score'>
In [2849]:
cluster_range = range(1, 30)  # Try 1 to 30 clusters
avg_distances = []

for k in cluster_range:
    kmeans = KMeans(n_clusters=k, random_state=42)
    kmeans.fit(input_pca)
    
    distances = np.min(kmeans.transform(input_pca), axis=1)
    avg_distance = np.mean(distances)
    avg_distances.append(avg_distance)

# Print results
for k, dist in zip(cluster_range, avg_distances):
    print(f"Clusters: {k}, Average distance: {dist:.2f}")

Clusters: 1, Average distance: 7.09
Clusters: 2, Average distance: 6.57
Clusters: 3, Average distance: 6.22
Clusters: 4, Average distance: 6.02
Clusters: 5, Average distance: 5.84
Clusters: 6, Average distance: 5.59
Clusters: 7, Average distance: 5.46
Clusters: 8, Average distance: 5.34
Clusters: 9, Average distance: 5.26
Clusters: 10, Average distance: 5.22
Clusters: 11, Average distance: 5.16
Clusters: 12, Average distance: 5.13
Clusters: 13, Average distance: 5.08
Clusters: 14, Average distance: 5.05
Clusters: 15, Average distance: 5.02
Clusters: 16, Average distance: 5.00
Clusters: 17, Average distance: 4.97
Clusters: 18, Average distance: 4.95
Clusters: 19, Average distance: 4.94
Clusters: 20, Average distance: 4.90
Clusters: 21, Average distance: 4.87
Clusters: 22, Average distance: 4.85
Clusters: 23, Average distance: 4.83
Clusters: 24, Average distance: 4.82
Clusters: 25, Average distance: 4.81
Clusters: 26, Average distance: 4.79
Clusters: 27, Average distance: 4.78
Clusters: 28, Average distance: 4.76
Clusters: 29, Average distance: 4.75
In [2850]:
cluster_range = range(1, 40)  # Try 1 to 40 clusters too
avg_distances = []

for k in cluster_range:
    kmeans = KMeans(n_clusters=k, random_state=42)
    kmeans.fit(input_pca)
    
    distances = np.min(kmeans.transform(input_pca), axis=1)
    avg_distance = np.mean(distances)
    avg_distances.append(avg_distance)

# Plot the results
plt.figure(figsize=(10, 6))
plt.plot(cluster_range, avg_distances, marker='o')
plt.xlabel('Number of Clusters')
plt.ylabel('Average Within-Cluster Distance')
plt.title('K-Means Clustering: Average Within-Cluster Distance')
plt.grid(True)
plt.show()
No description has been provided for this image
In [2851]:
plt.scatter(input_pca[:, 0], input_pca[:, 1], cmap='viridis')

centers = kmeans.cluster_centers_
plt.scatter(centers[:, 0], centers[:, 1], c='black', s=200, alpha=0.5);

/var/folders/y1/2cln1s193dgc6xy3yzry761r0000gn/T/ipykernel_845/3272324055.py:1: UserWarning: No data for colormapping provided via 'c'. Parameters 'cmap' will be ignored
  plt.scatter(input_pca[:, 0], input_pca[:, 1], cmap='viridis')
No description has been provided for this image

KNN Prediction on Population Data¶

In [2853]:
kmeans = KMeans(n_clusters=11, random_state=42)     #as advised by both yellow bricks' distortion score estimation
In [2854]:
model = kmeans.fit(input_pca)
pop_model_pred = model.predict(input_pca)
In [2862]:
pop_model_pred
Out[2862]:
array([ 0,  7, 10, ...,  4,  3,  9], dtype=int32)

Discussion 3.1: Apply Clustering to General Population¶

  • Our Elbow Plot gives roughly a 11 number of clusters to use really.
  • After reseaarching, we used yellowbrick's visualizer as read in this medium post.

In alll, we have gone for Yellow-bricks Elbow value because:

  • The automatic detection of the elbow point is a valuable feature that would require additional coding if done manually.
  • The inclusion of multiple metrics provides a more robust analysis

Step 3.2: Apply All Steps to the Customer Data¶

Now that you have clusters and cluster centers for the general population, it's time to see how the customer data maps on to those clusters. Take care to not confuse this for re-fitting all of the models to the customer data. Instead, you're going to use the fits from the general population to clean, transform, and cluster the customer data. In the last step of the project, you will interpret how the general population fits apply to the customer data.

  • Don't forget when loading in the customers data, that it is semicolon (;) delimited.
  • Apply the same feature wrangling, selection, and engineering steps to the customer demographics using the clean_data() function you created earlier. (You can assume that the customer demographics data has similar meaning behind missing data patterns as the general demographics data.)
  • Use the sklearn objects from the general demographics data, and apply their transformations to the customers data. That is, you should not be using a .fit() or .fit_transform() method to re-fit the old objects, nor should you be creating new sklearn objects! Carry the data through the feature scaling, PCA, and clustering steps, obtaining cluster assignments for all of the data in the customer demographics data.
In [2867]:
# Load in the customer demographics data.
customers = pd.read_csv('Udacity_CUSTOMERS_Subset.csv', sep=';')
In [2869]:
customers.head()
Out[2869]:
AGER_TYP ALTERSKATEGORIE_GROB ANREDE_KZ CJT_GESAMTTYP FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER FINANZTYP GEBURTSJAHR GFK_URLAUBERTYP GREEN_AVANTGARDE HEALTH_TYP LP_LEBENSPHASE_FEIN LP_LEBENSPHASE_GROB LP_FAMILIE_FEIN LP_FAMILIE_GROB LP_STATUS_FEIN LP_STATUS_GROB NATIONALITAET_KZ PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV SHOPPER_TYP SOHO_KZ TITEL_KZ VERS_TYP ZABEOTYP ALTER_HH ANZ_PERSONEN ANZ_TITEL HH_EINKOMMEN_SCORE KK_KUNDENTYP W_KEIT_KIND_HH WOHNDAUER_2008 ANZ_HAUSHALTE_AKTIV ANZ_HH_TITEL GEBAEUDETYP KONSUMNAEHE MIN_GEBAEUDEJAHR OST_WEST_KZ WOHNLAGE CAMEO_DEUG_2015 CAMEO_DEU_2015 CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_ANTG3 KBA05_ANTG4 KBA05_BAUMAX KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP KBA13_ANZAHL_PKW PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_BAUMAX PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB
0 2 4 1 5.0 5 1 5 1 2 2 2 0 4.0 1 1 20.0 5.0 2.0 2.0 10.0 5.0 1 4 5.0 6 5 2 6 6 7 3 4 1 3 1 1 2 1 3 0.0 0.0 1 3 10.0 2.0 0.0 1.0 NaN 6.0 9.0 1.0 0.0 1.0 5.0 1992.0 W 7.0 1 1A 13 2.0 2.0 0.0 0.0 0.0 4.0 3.0 2.0 4.0 4.0 1.0 4.0 3.0 1.0 1201.0 3.0 3.0 1.0 0.0 1.0 5.0 5.0 1.0 2.0 1.0
1 -1 4 1 NaN 5 1 5 1 3 2 2 0 NaN 0 1 NaN NaN NaN NaN NaN NaN 1 0 NaN 3 6 2 6 7 5 3 4 1 3 3 2 4 1 3 0.0 0.0 1 3 11.0 3.0 0.0 NaN NaN 0.0 9.0 NaN NaN NaN 5.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN
2 -1 4 2 2.0 5 1 5 1 4 4 2 0 3.0 1 2 13.0 3.0 1.0 1.0 10.0 5.0 1 4 5.0 2 2 1 3 3 7 7 1 2 7 5 6 4 1 1 0.0 0.0 2 3 6.0 1.0 0.0 1.0 NaN 6.0 9.0 1.0 0.0 8.0 1.0 1992.0 W 2.0 5 5D 34 2.0 2.0 0.0 0.0 0.0 3.0 7.0 4.0 1.0 3.0 3.0 3.0 1.0 7.0 433.0 2.0 3.0 3.0 1.0 3.0 3.0 2.0 3.0 5.0 3.0
3 1 4 1 2.0 5 1 5 2 1 2 6 0 10.0 0 2 0.0 0.0 0.0 0.0 9.0 4.0 1 1 3.0 6 5 3 4 7 5 3 4 3 3 3 3 3 4 0 0.0 0.0 1 1 8.0 0.0 0.0 4.0 NaN NaN 9.0 0.0 NaN 2.0 2.0 1992.0 W 7.0 4 4C 24 3.0 0.0 0.0 0.0 1.0 4.0 7.0 1.0 7.0 4.0 3.0 4.0 2.0 6.0 755.0 3.0 2.0 1.0 0.0 1.0 3.0 4.0 1.0 3.0 1.0
4 -1 3 1 6.0 3 1 4 4 5 2 2 1960 2.0 0 3 31.0 10.0 10.0 5.0 1.0 1.0 1 8 5.0 4 5 4 6 5 6 4 5 5 3 5 2 5 4 1 0.0 0.0 2 1 20.0 4.0 0.0 6.0 2.0 2.0 9.0 7.0 0.0 3.0 1.0 1992.0 W 3.0 7 7B 41 0.0 3.0 2.0 0.0 0.0 3.0 3.0 4.0 4.0 3.0 4.0 3.0 5.0 7.0 513.0 2.0 4.0 2.0 1.0 2.0 3.0 3.0 3.0 5.0 1.0
In [2871]:
customers.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 191652 entries, 0 to 191651
Data columns (total 85 columns):
 #   Column                 Non-Null Count   Dtype  
---  ------                 --------------   -----  
 0   AGER_TYP               191652 non-null  int64  
 1   ALTERSKATEGORIE_GROB   191652 non-null  int64  
 2   ANREDE_KZ              191652 non-null  int64  
 3   CJT_GESAMTTYP          188439 non-null  float64
 4   FINANZ_MINIMALIST      191652 non-null  int64  
 5   FINANZ_SPARER          191652 non-null  int64  
 6   FINANZ_VORSORGER       191652 non-null  int64  
 7   FINANZ_ANLEGER         191652 non-null  int64  
 8   FINANZ_UNAUFFAELLIGER  191652 non-null  int64  
 9   FINANZ_HAUSBAUER       191652 non-null  int64  
 10  FINANZTYP              191652 non-null  int64  
 11  GEBURTSJAHR            191652 non-null  int64  
 12  GFK_URLAUBERTYP        188439 non-null  float64
 13  GREEN_AVANTGARDE       191652 non-null  int64  
 14  HEALTH_TYP             191652 non-null  int64  
 15  LP_LEBENSPHASE_FEIN    188439 non-null  float64
 16  LP_LEBENSPHASE_GROB    188439 non-null  float64
 17  LP_FAMILIE_FEIN        188439 non-null  float64
 18  LP_FAMILIE_GROB        188439 non-null  float64
 19  LP_STATUS_FEIN         188439 non-null  float64
 20  LP_STATUS_GROB         188439 non-null  float64
 21  NATIONALITAET_KZ       191652 non-null  int64  
 22  PRAEGENDE_JUGENDJAHRE  191652 non-null  int64  
 23  RETOURTYP_BK_S         188439 non-null  float64
 24  SEMIO_SOZ              191652 non-null  int64  
 25  SEMIO_FAM              191652 non-null  int64  
 26  SEMIO_REL              191652 non-null  int64  
 27  SEMIO_MAT              191652 non-null  int64  
 28  SEMIO_VERT             191652 non-null  int64  
 29  SEMIO_LUST             191652 non-null  int64  
 30  SEMIO_ERL              191652 non-null  int64  
 31  SEMIO_KULT             191652 non-null  int64  
 32  SEMIO_RAT              191652 non-null  int64  
 33  SEMIO_KRIT             191652 non-null  int64  
 34  SEMIO_DOM              191652 non-null  int64  
 35  SEMIO_KAEM             191652 non-null  int64  
 36  SEMIO_PFLICHT          191652 non-null  int64  
 37  SEMIO_TRADV            191652 non-null  int64  
 38  SHOPPER_TYP            191652 non-null  int64  
 39  SOHO_KZ                145056 non-null  float64
 40  TITEL_KZ               145056 non-null  float64
 41  VERS_TYP               191652 non-null  int64  
 42  ZABEOTYP               191652 non-null  int64  
 43  ALTER_HH               145056 non-null  float64
 44  ANZ_PERSONEN           145056 non-null  float64
 45  ANZ_TITEL              145056 non-null  float64
 46  HH_EINKOMMEN_SCORE     188684 non-null  float64
 47  KK_KUNDENTYP           79715 non-null   float64
 48  W_KEIT_KIND_HH         137910 non-null  float64
 49  WOHNDAUER_2008         145056 non-null  float64
 50  ANZ_HAUSHALTE_AKTIV    141725 non-null  float64
 51  ANZ_HH_TITEL           139542 non-null  float64
 52  GEBAEUDETYP            141725 non-null  float64
 53  KONSUMNAEHE            145001 non-null  float64
 54  MIN_GEBAEUDEJAHR       141725 non-null  float64
 55  OST_WEST_KZ            141725 non-null  object 
 56  WOHNLAGE               141725 non-null  float64
 57  CAMEO_DEUG_2015        141224 non-null  object 
 58  CAMEO_DEU_2015         141224 non-null  object 
 59  CAMEO_INTL_2015        141224 non-null  object 
 60  KBA05_ANTG1            135672 non-null  float64
 61  KBA05_ANTG2            135672 non-null  float64
 62  KBA05_ANTG3            135672 non-null  float64
 63  KBA05_ANTG4            135672 non-null  float64
 64  KBA05_BAUMAX           135672 non-null  float64
 65  KBA05_GBZ              135672 non-null  float64
 66  BALLRAUM               141693 non-null  float64
 67  EWDICHTE               141693 non-null  float64
 68  INNENSTADT             141693 non-null  float64
 69  GEBAEUDETYP_RASTER     141725 non-null  float64
 70  KKK                    137392 non-null  float64
 71  MOBI_REGIO             135672 non-null  float64
 72  ONLINE_AFFINITAET      188439 non-null  float64
 73  REGIOTYP               137392 non-null  float64
 74  KBA13_ANZAHL_PKW       140371 non-null  float64
 75  PLZ8_ANTG1             138888 non-null  float64
 76  PLZ8_ANTG2             138888 non-null  float64
 77  PLZ8_ANTG3             138888 non-null  float64
 78  PLZ8_ANTG4             138888 non-null  float64
 79  PLZ8_BAUMAX            138888 non-null  float64
 80  PLZ8_HHZ               138888 non-null  float64
 81  PLZ8_GBZ               138888 non-null  float64
 82  ARBEIT                 141176 non-null  float64
 83  ORTSGR_KLS9            141176 non-null  float64
 84  RELAT_AB               141176 non-null  float64
dtypes: float64(49), int64(32), object(4)
memory usage: 124.3+ MB
In [2873]:
customers.isna().sum()
Out[2873]:
AGER_TYP                      0
ALTERSKATEGORIE_GROB          0
ANREDE_KZ                     0
CJT_GESAMTTYP              3213
FINANZ_MINIMALIST             0
FINANZ_SPARER                 0
FINANZ_VORSORGER              0
FINANZ_ANLEGER                0
FINANZ_UNAUFFAELLIGER         0
FINANZ_HAUSBAUER              0
FINANZTYP                     0
GEBURTSJAHR                   0
GFK_URLAUBERTYP            3213
GREEN_AVANTGARDE              0
HEALTH_TYP                    0
LP_LEBENSPHASE_FEIN        3213
LP_LEBENSPHASE_GROB        3213
LP_FAMILIE_FEIN            3213
LP_FAMILIE_GROB            3213
LP_STATUS_FEIN             3213
LP_STATUS_GROB             3213
NATIONALITAET_KZ              0
PRAEGENDE_JUGENDJAHRE         0
RETOURTYP_BK_S             3213
SEMIO_SOZ                     0
SEMIO_FAM                     0
SEMIO_REL                     0
SEMIO_MAT                     0
SEMIO_VERT                    0
SEMIO_LUST                    0
SEMIO_ERL                     0
SEMIO_KULT                    0
SEMIO_RAT                     0
SEMIO_KRIT                    0
SEMIO_DOM                     0
SEMIO_KAEM                    0
SEMIO_PFLICHT                 0
SEMIO_TRADV                   0
SHOPPER_TYP                   0
SOHO_KZ                   46596
TITEL_KZ                  46596
VERS_TYP                      0
ZABEOTYP                      0
ALTER_HH                  46596
ANZ_PERSONEN              46596
ANZ_TITEL                 46596
HH_EINKOMMEN_SCORE         2968
KK_KUNDENTYP             111937
W_KEIT_KIND_HH            53742
WOHNDAUER_2008            46596
ANZ_HAUSHALTE_AKTIV       49927
ANZ_HH_TITEL              52110
GEBAEUDETYP               49927
KONSUMNAEHE               46651
MIN_GEBAEUDEJAHR          49927
OST_WEST_KZ               49927
WOHNLAGE                  49927
CAMEO_DEUG_2015           50428
CAMEO_DEU_2015            50428
CAMEO_INTL_2015           50428
KBA05_ANTG1               55980
KBA05_ANTG2               55980
KBA05_ANTG3               55980
KBA05_ANTG4               55980
KBA05_BAUMAX              55980
KBA05_GBZ                 55980
BALLRAUM                  49959
EWDICHTE                  49959
INNENSTADT                49959
GEBAEUDETYP_RASTER        49927
KKK                       54260
MOBI_REGIO                55980
ONLINE_AFFINITAET          3213
REGIOTYP                  54260
KBA13_ANZAHL_PKW          51281
PLZ8_ANTG1                52764
PLZ8_ANTG2                52764
PLZ8_ANTG3                52764
PLZ8_ANTG4                52764
PLZ8_BAUMAX               52764
PLZ8_HHZ                  52764
PLZ8_GBZ                  52764
ARBEIT                    50476
ORTSGR_KLS9               50476
RELAT_AB                  50476
dtype: int64

Preprocessing the Customers Data¶

In [2876]:
cust = clean_data(customers)           #using cleaning function
In [2878]:
cust.head()
Out[2878]:
ALTERSKATEGORIE_GROB ANREDE_KZ FINANZ_MINIMALIST FINANZ_SPARER FINANZ_VORSORGER FINANZ_ANLEGER FINANZ_UNAUFFAELLIGER FINANZ_HAUSBAUER GREEN_AVANTGARDE HEALTH_TYP PRAEGENDE_JUGENDJAHRE RETOURTYP_BK_S SEMIO_SOZ SEMIO_FAM SEMIO_REL SEMIO_MAT SEMIO_VERT SEMIO_LUST SEMIO_ERL SEMIO_KULT SEMIO_RAT SEMIO_KRIT SEMIO_DOM SEMIO_KAEM SEMIO_PFLICHT SEMIO_TRADV VERS_TYP HH_EINKOMMEN_SCORE W_KEIT_KIND_HH KONSUMNAEHE CAMEO_INTL_2015 KBA05_ANTG1 KBA05_ANTG2 KBA05_GBZ BALLRAUM EWDICHTE INNENSTADT GEBAEUDETYP_RASTER KKK MOBI_REGIO ONLINE_AFFINITAET REGIOTYP PLZ8_ANTG1 PLZ8_ANTG2 PLZ8_ANTG3 PLZ8_ANTG4 PLZ8_HHZ PLZ8_GBZ ARBEIT ORTSGR_KLS9 RELAT_AB wealth life_stage generation movement
0 4.0 1 5 1 5 1 2 2 1 1.0 4.0 5.0 6 5 2 6 6 7 3 4 1 3 1 1 2 1 1.0 1.0 6.0 5.0 13.0 2.0 2.0 4.0 3.0 2.0 4.0 4.0 1.0 4.0 3.0 1.0 3.0 3.0 1.0 0.0 5.0 5.0 1.0 2.0 1.0 0 2 1 1
1 4.0 1 5 1 5 1 3 2 0 1.0 NaN NaN 3 6 2 6 7 5 3 4 1 3 3 2 4 1 1.0 NaN NaN 5.0 NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN 4 4 5 1
2 4.0 2 5 1 5 1 4 4 1 2.0 4.0 5.0 2 2 1 3 3 7 7 1 2 7 5 6 4 1 2.0 1.0 6.0 1.0 34.0 2.0 2.0 3.0 7.0 4.0 1.0 3.0 3.0 3.0 1.0 7.0 2.0 3.0 3.0 1.0 3.0 2.0 3.0 5.0 3.0 2 3 1 1
3 4.0 1 5 1 5 2 1 2 0 2.0 1.0 3.0 6 5 3 4 7 5 3 4 3 3 3 3 3 4 1.0 4.0 NaN 2.0 24.0 3.0 0.0 4.0 7.0 1.0 7.0 4.0 3.0 4.0 2.0 6.0 3.0 2.0 1.0 0.0 3.0 4.0 1.0 3.0 1.0 1 3 0 0
4 3.0 1 3 1 4 4 5 2 0 3.0 8.0 5.0 4 5 4 6 5 6 4 5 5 3 5 2 5 4 2.0 6.0 2.0 1.0 41.0 0.0 3.0 3.0 3.0 4.0 4.0 3.0 4.0 3.0 5.0 7.0 2.0 4.0 2.0 1.0 3.0 3.0 3.0 5.0 1.0 3 0 3 0
In [2880]:
cust.shape
Out[2880]:
(191652, 55)
In [2882]:
cust.isna().sum()
Out[2882]:
ALTERSKATEGORIE_GROB       240
ANREDE_KZ                    0
FINANZ_MINIMALIST            0
FINANZ_SPARER                0
FINANZ_VORSORGER             0
FINANZ_ANLEGER               0
FINANZ_UNAUFFAELLIGER        0
FINANZ_HAUSBAUER             0
GREEN_AVANTGARDE             0
HEALTH_TYP               48990
PRAEGENDE_JUGENDJAHRE    48487
RETOURTYP_BK_S            3213
SEMIO_SOZ                    0
SEMIO_FAM                    0
SEMIO_REL                    0
SEMIO_MAT                    0
SEMIO_VERT                   0
SEMIO_LUST                   0
SEMIO_ERL                    0
SEMIO_KULT                   0
SEMIO_RAT                    0
SEMIO_KRIT                   0
SEMIO_DOM                    0
SEMIO_KAEM                   0
SEMIO_PFLICHT                0
SEMIO_TRADV                  0
VERS_TYP                 48990
HH_EINKOMMEN_SCORE        2968
W_KEIT_KIND_HH           56937
KONSUMNAEHE              46651
CAMEO_INTL_2015          50554
KBA05_ANTG1              55980
KBA05_ANTG2              55980
KBA05_GBZ                55982
BALLRAUM                 49959
EWDICHTE                 49959
INNENSTADT               49959
GEBAEUDETYP_RASTER       49927
KKK                      60064
MOBI_REGIO               55980
ONLINE_AFFINITAET         3213
REGIOTYP                 60064
PLZ8_ANTG1               52764
PLZ8_ANTG2               52764
PLZ8_ANTG3               52764
PLZ8_ANTG4               52764
PLZ8_HHZ                 52764
PLZ8_GBZ                 52764
ARBEIT                   50499
ORTSGR_KLS9              50476
RELAT_AB                 50499
wealth                       0
life_stage                   0
generation                   0
movement                     0
dtype: int64
In [2884]:
def add_missing_dummy_columns(d, columns):
    missing_cols = set(columns) - set(d.columns)
    print(missing_cols)
    for c in missing_cols:
        d[c] = 0
           
def fix_columns(d, columns):  
    add_missing_dummy_columns(d, columns)
    # make sure we have all the columns we need
    assert(set(columns) - set(d.columns) == set())
    extra_cols = set(d.columns) - set(columns)
    if extra_cols:
        print("extra columns:", extra_cols)
    d = d[columns]
    return d
customers = fix_columns(cust, df.columns.tolist())

set()
extra columns: {'GREEN_AVANTGARDE'}
In [2886]:
cust = cust.drop(['GREEN_AVANTGARDE'], axis=1)
In [2985]:
#inputing nan values as in the Population Data


cust_imp = pd.DataFrame(imp_median.transform(cust), columns=cust.columns)
In [2987]:
#verifying nan count

cust_imp.isna().sum().sum()
Out[2987]:
0
In [2989]:
#standerdizing using scaler predefined earlier

cust_scaled = pd.DataFrame(scaler.transform(cust_imp), columns=cust_imp.columns)
In [2991]:
#reconfirming PCA value
pca?

Type:        PCA
String form: PCA(n_components=36)
File:        ~/opt/anaconda3/lib/python3.11/site-packages/sklearn/decomposition/_pca.py
Docstring:  
Principal component analysis (PCA).

Linear dimensionality reduction using Singular Value Decomposition of the
data to project it to a lower dimensional space. The input data is centered
but not scaled for each feature before applying the SVD.

It uses the LAPACK implementation of the full SVD or a randomized truncated
SVD by the method of Halko et al. 2009, depending on the shape of the input
data and the number of components to extract.

With sparse inputs, the ARPACK implementation of the truncated SVD can be
used (i.e. through :func:`scipy.sparse.linalg.svds`). Alternatively, one
may consider :class:`TruncatedSVD` where the data are not centered.

Notice that this class only supports sparse inputs for some solvers such as
"arpack" and "covariance_eigh". See :class:`TruncatedSVD` for an
alternative with sparse data.

For a usage example, see
:ref:`sphx_glr_auto_examples_decomposition_plot_pca_iris.py`

Read more in the :ref:`User Guide <PCA>`.

Parameters
----------
n_components : int, float or 'mle', default=None
    Number of components to keep.
    if n_components is not set all components are kept::

        n_components == min(n_samples, n_features)

    If ``n_components == 'mle'`` and ``svd_solver == 'full'``, Minka's
    MLE is used to guess the dimension. Use of ``n_components == 'mle'``
    will interpret ``svd_solver == 'auto'`` as ``svd_solver == 'full'``.

    If ``0 < n_components < 1`` and ``svd_solver == 'full'``, select the
    number of components such that the amount of variance that needs to be
    explained is greater than the percentage specified by n_components.

    If ``svd_solver == 'arpack'``, the number of components must be
    strictly less than the minimum of n_features and n_samples.

    Hence, the None case results in::

        n_components == min(n_samples, n_features) - 1

copy : bool, default=True
    If False, data passed to fit are overwritten and running
    fit(X).transform(X) will not yield the expected results,
    use fit_transform(X) instead.

whiten : bool, default=False
    When True (False by default) the `components_` vectors are multiplied
    by the square root of n_samples and then divided by the singular values
    to ensure uncorrelated outputs with unit component-wise variances.

    Whitening will remove some information from the transformed signal
    (the relative variance scales of the components) but can sometime
    improve the predictive accuracy of the downstream estimators by
    making their data respect some hard-wired assumptions.

svd_solver : {'auto', 'full', 'covariance_eigh', 'arpack', 'randomized'},            default='auto'
    "auto" :
        The solver is selected by a default 'auto' policy is based on `X.shape` and
        `n_components`: if the input data has fewer than 1000 features and
        more than 10 times as many samples, then the "covariance_eigh"
        solver is used. Otherwise, if the input data is larger than 500x500
        and the number of components to extract is lower than 80% of the
        smallest dimension of the data, then the more efficient
        "randomized" method is selected. Otherwise the exact "full" SVD is
        computed and optionally truncated afterwards.
    "full" :
        Run exact full SVD calling the standard LAPACK solver via
        `scipy.linalg.svd` and select the components by postprocessing
    "covariance_eigh" :
        Precompute the covariance matrix (on centered data), run a
        classical eigenvalue decomposition on the covariance matrix
        typically using LAPACK and select the components by postprocessing.
        This solver is very efficient for n_samples >> n_features and small
        n_features. It is, however, not tractable otherwise for large
        n_features (large memory footprint required to materialize the
        covariance matrix). Also note that compared to the "full" solver,
        this solver effectively doubles the condition number and is
        therefore less numerical stable (e.g. on input data with a large
        range of singular values).
    "arpack" :
        Run SVD truncated to `n_components` calling ARPACK solver via
        `scipy.sparse.linalg.svds`. It requires strictly
        `0 < n_components < min(X.shape)`
    "randomized" :
        Run randomized SVD by the method of Halko et al.

    .. versionadded:: 0.18.0

    .. versionchanged:: 1.5
        Added the 'covariance_eigh' solver.

tol : float, default=0.0
    Tolerance for singular values computed by svd_solver == 'arpack'.
    Must be of range [0.0, infinity).

    .. versionadded:: 0.18.0

iterated_power : int or 'auto', default='auto'
    Number of iterations for the power method computed by
    svd_solver == 'randomized'.
    Must be of range [0, infinity).

    .. versionadded:: 0.18.0

n_oversamples : int, default=10
    This parameter is only relevant when `svd_solver="randomized"`.
    It corresponds to the additional number of random vectors to sample the
    range of `X` so as to ensure proper conditioning. See
    :func:`~sklearn.utils.extmath.randomized_svd` for more details.

    .. versionadded:: 1.1

power_iteration_normalizer : {'auto', 'QR', 'LU', 'none'}, default='auto'
    Power iteration normalizer for randomized SVD solver.
    Not used by ARPACK. See :func:`~sklearn.utils.extmath.randomized_svd`
    for more details.

    .. versionadded:: 1.1

random_state : int, RandomState instance or None, default=None
    Used when the 'arpack' or 'randomized' solvers are used. Pass an int
    for reproducible results across multiple function calls.
    See :term:`Glossary <random_state>`.

    .. versionadded:: 0.18.0

Attributes
----------
components_ : ndarray of shape (n_components, n_features)
    Principal axes in feature space, representing the directions of
    maximum variance in the data. Equivalently, the right singular
    vectors of the centered input data, parallel to its eigenvectors.
    The components are sorted by decreasing ``explained_variance_``.

explained_variance_ : ndarray of shape (n_components,)
    The amount of variance explained by each of the selected components.
    The variance estimation uses `n_samples - 1` degrees of freedom.

    Equal to n_components largest eigenvalues
    of the covariance matrix of X.

    .. versionadded:: 0.18

explained_variance_ratio_ : ndarray of shape (n_components,)
    Percentage of variance explained by each of the selected components.

    If ``n_components`` is not set then all components are stored and the
    sum of the ratios is equal to 1.0.

singular_values_ : ndarray of shape (n_components,)
    The singular values corresponding to each of the selected components.
    The singular values are equal to the 2-norms of the ``n_components``
    variables in the lower-dimensional space.

    .. versionadded:: 0.19

mean_ : ndarray of shape (n_features,)
    Per-feature empirical mean, estimated from the training set.

    Equal to `X.mean(axis=0)`.

n_components_ : int
    The estimated number of components. When n_components is set
    to 'mle' or a number between 0 and 1 (with svd_solver == 'full') this
    number is estimated from input data. Otherwise it equals the parameter
    n_components, or the lesser value of n_features and n_samples
    if n_components is None.

n_samples_ : int
    Number of samples in the training data.

noise_variance_ : float
    The estimated noise covariance following the Probabilistic PCA model
    from Tipping and Bishop 1999. See "Pattern Recognition and
    Machine Learning" by C. Bishop, 12.2.1 p. 574 or
    http://www.miketipping.com/papers/met-mppca.pdf. It is required to
    compute the estimated data covariance and score samples.

    Equal to the average of (min(n_features, n_samples) - n_components)
    smallest eigenvalues of the covariance matrix of X.

n_features_in_ : int
    Number of features seen during :term:`fit`.

    .. versionadded:: 0.24

feature_names_in_ : ndarray of shape (`n_features_in_`,)
    Names of features seen during :term:`fit`. Defined only when `X`
    has feature names that are all strings.

    .. versionadded:: 1.0

See Also
--------
KernelPCA : Kernel Principal Component Analysis.
SparsePCA : Sparse Principal Component Analysis.
TruncatedSVD : Dimensionality reduction using truncated SVD.
IncrementalPCA : Incremental Principal Component Analysis.

References
----------
For n_components == 'mle', this class uses the method from:
`Minka, T. P.. "Automatic choice of dimensionality for PCA".
In NIPS, pp. 598-604 <https://tminka.github.io/papers/pca/minka-pca.pdf>`_

Implements the probabilistic PCA model from:
`Tipping, M. E., and Bishop, C. M. (1999). "Probabilistic principal
component analysis". Journal of the Royal Statistical Society:
Series B (Statistical Methodology), 61(3), 611-622.
<http://www.miketipping.com/papers/met-mppca.pdf>`_
via the score and score_samples methods.

For svd_solver == 'arpack', refer to `scipy.sparse.linalg.svds`.

For svd_solver == 'randomized', see:
:doi:`Halko, N., Martinsson, P. G., and Tropp, J. A. (2011).
"Finding structure with randomness: Probabilistic algorithms for
constructing approximate matrix decompositions".
SIAM review, 53(2), 217-288.
<10.1137/090771806>`
and also
:doi:`Martinsson, P. G., Rokhlin, V., and Tygert, M. (2011).
"A randomized algorithm for the decomposition of matrices".
Applied and Computational Harmonic Analysis, 30(1), 47-68.
<10.1016/j.acha.2010.02.003>`

Examples
--------
>>> import numpy as np
>>> from sklearn.decomposition import PCA
>>> X = np.array([[-1, -1], [-2, -1], [-3, -2], [1, 1], [2, 1], [3, 2]])
>>> pca = PCA(n_components=2)
>>> pca.fit(X)
PCA(n_components=2)
>>> print(pca.explained_variance_ratio_)
[0.9924... 0.0075...]
>>> print(pca.singular_values_)
[6.30061... 0.54980...]

>>> pca = PCA(n_components=2, svd_solver='full')
>>> pca.fit(X)
PCA(n_components=2, svd_solver='full')
>>> print(pca.explained_variance_ratio_)
[0.9924... 0.00755...]
>>> print(pca.singular_values_)
[6.30061... 0.54980...]

>>> pca = PCA(n_components=1, svd_solver='arpack')
>>> pca.fit(X)
PCA(n_components=1, svd_solver='arpack')
>>> print(pca.explained_variance_ratio_)
[0.99244...]
>>> print(pca.singular_values_)
[6.30061...]
In [2993]:
#applying pca now

cust_pca = pca.transform(cust_scaled)

/Users/chrismo/opt/anaconda3/lib/python3.11/site-packages/sklearn/base.py:486: UserWarning: X has feature names, but PCA was fitted without feature names
  warnings.warn(

Applying Clustering from General Population Data unto the Customer Data¶

In [2996]:
model?

Type:        KMeans
String form: KMeans(n_clusters=11, random_state=42)
File:        ~/opt/anaconda3/lib/python3.11/site-packages/sklearn/cluster/_kmeans.py
Docstring:  
K-Means clustering.

Read more in the :ref:`User Guide <k_means>`.

Parameters
----------

n_clusters : int, default=8
    The number of clusters to form as well as the number of
    centroids to generate.

    For an example of how to choose an optimal value for `n_clusters` refer to
    :ref:`sphx_glr_auto_examples_cluster_plot_kmeans_silhouette_analysis.py`.

init : {'k-means++', 'random'}, callable or array-like of shape             (n_clusters, n_features), default='k-means++'
    Method for initialization:

    * 'k-means++' : selects initial cluster centroids using sampling             based on an empirical probability distribution of the points'             contribution to the overall inertia. This technique speeds up             convergence. The algorithm implemented is "greedy k-means++". It             differs from the vanilla k-means++ by making several trials at             each sampling step and choosing the best centroid among them.

    * 'random': choose `n_clusters` observations (rows) at random from         data for the initial centroids.

    * If an array is passed, it should be of shape (n_clusters, n_features)        and gives the initial centers.

    * If a callable is passed, it should take arguments X, n_clusters and a        random state and return an initialization.

    For an example of how to use the different `init` strategy, see the example
    entitled :ref:`sphx_glr_auto_examples_cluster_plot_kmeans_digits.py`.

n_init : 'auto' or int, default='auto'
    Number of times the k-means algorithm is run with different centroid
    seeds. The final results is the best output of `n_init` consecutive runs
    in terms of inertia. Several runs are recommended for sparse
    high-dimensional problems (see :ref:`kmeans_sparse_high_dim`).

    When `n_init='auto'`, the number of runs depends on the value of init:
    10 if using `init='random'` or `init` is a callable;
    1 if using `init='k-means++'` or `init` is an array-like.

    .. versionadded:: 1.2
       Added 'auto' option for `n_init`.

    .. versionchanged:: 1.4
       Default value for `n_init` changed to `'auto'`.

max_iter : int, default=300
    Maximum number of iterations of the k-means algorithm for a
    single run.

tol : float, default=1e-4
    Relative tolerance with regards to Frobenius norm of the difference
    in the cluster centers of two consecutive iterations to declare
    convergence.

verbose : int, default=0
    Verbosity mode.

random_state : int, RandomState instance or None, default=None
    Determines random number generation for centroid initialization. Use
    an int to make the randomness deterministic.
    See :term:`Glossary <random_state>`.

copy_x : bool, default=True
    When pre-computing distances it is more numerically accurate to center
    the data first. If copy_x is True (default), then the original data is
    not modified. If False, the original data is modified, and put back
    before the function returns, but small numerical differences may be
    introduced by subtracting and then adding the data mean. Note that if
    the original data is not C-contiguous, a copy will be made even if
    copy_x is False. If the original data is sparse, but not in CSR format,
    a copy will be made even if copy_x is False.

algorithm : {"lloyd", "elkan"}, default="lloyd"
    K-means algorithm to use. The classical EM-style algorithm is `"lloyd"`.
    The `"elkan"` variation can be more efficient on some datasets with
    well-defined clusters, by using the triangle inequality. However it's
    more memory intensive due to the allocation of an extra array of shape
    `(n_samples, n_clusters)`.

    .. versionchanged:: 0.18
        Added Elkan algorithm

    .. versionchanged:: 1.1
        Renamed "full" to "lloyd", and deprecated "auto" and "full".
        Changed "auto" to use "lloyd" instead of "elkan".

Attributes
----------
cluster_centers_ : ndarray of shape (n_clusters, n_features)
    Coordinates of cluster centers. If the algorithm stops before fully
    converging (see ``tol`` and ``max_iter``), these will not be
    consistent with ``labels_``.

labels_ : ndarray of shape (n_samples,)
    Labels of each point

inertia_ : float
    Sum of squared distances of samples to their closest cluster center,
    weighted by the sample weights if provided.

n_iter_ : int
    Number of iterations run.

n_features_in_ : int
    Number of features seen during :term:`fit`.

    .. versionadded:: 0.24

feature_names_in_ : ndarray of shape (`n_features_in_`,)
    Names of features seen during :term:`fit`. Defined only when `X`
    has feature names that are all strings.

    .. versionadded:: 1.0

See Also
--------
MiniBatchKMeans : Alternative online implementation that does incremental
    updates of the centers positions using mini-batches.
    For large scale learning (say n_samples > 10k) MiniBatchKMeans is
    probably much faster than the default batch implementation.

Notes
-----
The k-means problem is solved using either Lloyd's or Elkan's algorithm.

The average complexity is given by O(k n T), where n is the number of
samples and T is the number of iteration.

The worst case complexity is given by O(n^(k+2/p)) with
n = n_samples, p = n_features.
Refer to :doi:`"How slow is the k-means method?" D. Arthur and S. Vassilvitskii -
SoCG2006.<10.1145/1137856.1137880>` for more details.

In practice, the k-means algorithm is very fast (one of the fastest
clustering algorithms available), but it falls in local minima. That's why
it can be useful to restart it several times.

If the algorithm stops before fully converging (because of ``tol`` or
``max_iter``), ``labels_`` and ``cluster_centers_`` will not be consistent,
i.e. the ``cluster_centers_`` will not be the means of the points in each
cluster. Also, the estimator will reassign ``labels_`` after the last
iteration to make ``labels_`` consistent with ``predict`` on the training
set.

Examples
--------

>>> from sklearn.cluster import KMeans
>>> import numpy as np
>>> X = np.array([[1, 2], [1, 4], [1, 0],
...               [10, 2], [10, 4], [10, 0]])
>>> kmeans = KMeans(n_clusters=2, random_state=0, n_init="auto").fit(X)
>>> kmeans.labels_
array([1, 1, 1, 0, 0, 0], dtype=int32)
>>> kmeans.predict([[0, 0], [12, 3]])
array([1, 0], dtype=int32)
>>> kmeans.cluster_centers_
array([[10.,  2.],
       [ 1.,  2.]])

For a more detailed example of K-Means using the iris dataset see
:ref:`sphx_glr_auto_examples_cluster_plot_cluster_iris.py`.

For examples of common problems with K-Means and how to address them see
:ref:`sphx_glr_auto_examples_cluster_plot_kmeans_assumptions.py`.

For an example of how to use K-Means to perform color quantization see
:ref:`sphx_glr_auto_examples_cluster_plot_color_quantization.py`.

For a demonstration of how K-Means can be used to cluster text documents see
:ref:`sphx_glr_auto_examples_text_plot_document_clustering.py`.

For a comparison between K-Means and MiniBatchKMeans refer to example
:ref:`sphx_glr_auto_examples_cluster_plot_mini_batch_kmeans.py`.
In [2998]:
cust_model_pred = model.predict(cust_pca)         #same model that trained our population df
In [3000]:
cust_model_pred
Out[3000]:
array([8, 9, 6, ..., 8, 6, 1], dtype=int32)
In [2962]:
# Apply preprocessing, feature transformation, and clustering from the general
# demographics onto the customer data, obtaining cluster predictions for the
# customer demographics data.

Step 3.3: Compare Customer Data to Demographics Data¶

At this point, you have clustered data based on demographics of the general population of Germany, and seen how the customer data for a mail-order sales company maps onto those demographic clusters. In this final substep, you will compare the two cluster distributions to see where the strongest customer base for the company is.

Consider the proportion of persons in each cluster for the general population, and the proportions for the customers. If we think the company's customer base to be universal, then the cluster assignment proportions should be fairly similar between the two. If there are only particular segments of the population that are interested in the company's products, then we should see a mismatch from one to the other. If there is a higher proportion of persons in a cluster for the customer data compared to the general population (e.g. 5% of persons are assigned to a cluster for the general population, but 15% of the customer data is closest to that cluster's centroid) then that suggests the people in that cluster to be a target audience for the company. On the other hand, the proportion of the data in a cluster being larger in the general population than the customer data (e.g. only 2% of customers closest to a population centroid that captures 6% of the data) suggests that group of persons to be outside of the target demographics.

Take a look at the following points in this step:

  • Compute the proportion of data points in each cluster for the general population and the customer data. Visualizations will be useful here: both for the individual dataset proportions, but also to visualize the ratios in cluster representation between groups. Seaborn's countplot() or barplot() function could be handy.
    • Recall the analysis you performed in step 1.1.3 of the project, where you separated out certain data points from the dataset if they had more than a specified threshold of missing values. If you found that this group was qualitatively different from the main bulk of the data, you should treat this as an additional data cluster in this analysis. Make sure that you account for the number of data points in this subset, for both the general population and customer datasets, when making your computations!
  • Which cluster or clusters are overrepresented in the customer dataset compared to the general population? Select at least one such cluster and infer what kind of people might be represented by that cluster. Use the principal component interpretations from step 2.3 or look at additional components to help you make this inference. Alternatively, you can use the .inverse_transform() method of the PCA and StandardScaler objects to transform centroids back to the original data space and interpret the retrieved values directly.
  • Perform a similar investigation for the underrepresented clusters. Which cluster or clusters are underrepresented in the customer dataset compared to the general population, and what kinds of people are typified by these clusters?
In [2965]:
# Define the desired clusters
desired_clusters = [0,1,2,3,4,5,6,7,8,9,10,11]  # Add up to 11 clusters
In [2967]:
# Filter data to include only desired clusters
cust__pred = pd.Series(cust_model_pred)
pop__pred = pd.Series(pop_model_pred)
In [2969]:
# Compare the proportion of data in each cluster for the customer data to the
# proportion of data in each cluster for the general population.

# Create the figure and subplots
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 6))

# Plot customer data
sns.countplot(x=cust_model_pred, palette='Set2', ax=ax1, hue=cust_model_pred, legend=False)
ax1.set_title('Customer Population Data')
ax1.set_xlabel('Cluster')
ax1.set_ylabel('Count')

# Plot general demographic data
sns.countplot(x=pop_model_pred, palette='Set2', ax=ax2, hue=pop_model_pred, legend=False)
ax2.set_title('General Population Data')
ax2.set_xlabel('Cluster')
ax2.set_ylabel('Count')

# Adjust layout and display the plot
plt.tight_layout()
plt.show();
No description has been provided for this image
In [2970]:
# Create the figure
plt.figure(figsize=(10, 6))

# Plot customer data
sns.countplot(x=cust_model_pred, palette='Set3', order=desired_clusters, label='Customer Data', alpha=0.6)

# Plot general demographic data
sns.countplot(x=pop_model_pred, palette='Set2', order=desired_clusters, label='General Demographic', alpha=0.6)

# Add labels and title
plt.xlabel('Cluster')
plt.ylabel('Count')
plt.title('Overlap of Customer Data and then Population Data')


# Display the plot
plt.tight_layout()
plt.show();

/var/folders/y1/2cln1s193dgc6xy3yzry761r0000gn/T/ipykernel_845/2898750846.py:5: FutureWarning: 

Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect.

  sns.countplot(x=cust_model_pred, palette='Set3', order=desired_clusters, label='Customer Data', alpha=0.6)
/var/folders/y1/2cln1s193dgc6xy3yzry761r0000gn/T/ipykernel_845/2898750846.py:8: FutureWarning: 

Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect.

  sns.countplot(x=pop_model_pred, palette='Set2', order=desired_clusters, label='General Demographic', alpha=0.6)
No description has been provided for this image
In [2973]:
# Generating proportion for the customer clusters
labels, values = zip(*Counter(cust_model_pred).items())
v = list(values)
v[:] = [x/len(cust_model_pred) for x in v]
indexes = np.arange(len(labels))

# Generating proportion for the azdias clusters
labels1, values1 = zip(*Counter(pop_model_pred).items())
v1 = list(values1)
v1[:] = [x/len(pop_model_pred) for x in v1]

# Configuring the graph
width = 0.35  # Width of each bar
spacing = 0.2  # Space between clusters
plt.figure(figsize=(12, 8))

# Calculate new x-coordinates for grouped bars
x = np.arange(0, len(labels) * (2*width + spacing), 2*width + spacing)

# Plot bars
plt.bar(x, v1, width, label='Population Data')
plt.bar(x + width, v, width, color='r', label='Customer Data')

# Adjust x-axis ticks and labels
plt.xticks(x + width/2, labels)

plt.legend(loc='best')
plt.xlabel('Cluster')
plt.ylabel('% of Total Customers')
plt.grid()

plt.show()
No description has been provided for this image

From the looks of it really, the Overrepresented bits are the 0, 8th and then the 9th clusters respectively¶

In [2976]:
# What kinds of people are part of a cluster that is overrepresented in the
# customer data compared to the general population?
# Exploring the 0th cluster now 


cluster_0 = model.cluster_centers_[0]
pca_0 = pca.inverse_transform(cluster_0)
pca_0_2d = np.array(pca_0).reshape(1, -1)  # Reshape to 2D
overrepresented = scaler.inverse_transform(pca_0_2d)

overrepresented = pd.Series(data=overrepresented.flatten(), index=list(df.keys()))
display(overrepresented.head(45))

ALTERSKATEGORIE_GROB      2.643979
ANREDE_KZ                 1.432840
FINANZ_MINIMALIST         3.791299
FINANZ_SPARER             3.165392
FINANZ_VORSORGER          3.888655
FINANZ_ANLEGER            4.629506
FINANZ_UNAUFFAELLIGER     4.793987
FINANZ_HAUSBAUER          2.779347
HEALTH_TYP                1.803794
PRAEGENDE_JUGENDJAHRE     6.630160
RETOURTYP_BK_S            3.621399
SEMIO_SOZ                 2.084913
SEMIO_FAM                 5.961617
SEMIO_REL                 6.840431
SEMIO_MAT                 4.830767
SEMIO_VERT                0.967002
SEMIO_LUST                5.912300
SEMIO_ERL                 3.181842
SEMIO_KULT                2.933584
SEMIO_RAT                 3.303255
SEMIO_KRIT                7.044412
SEMIO_DOM                 5.636096
SEMIO_KAEM                5.587260
SEMIO_PFLICHT             4.245575
SEMIO_TRADV               2.413348
VERS_TYP                  2.032865
HH_EINKOMMEN_SCORE        1.074767
W_KEIT_KIND_HH            4.056209
KONSUMNAEHE               3.055724
CAMEO_INTL_2015          37.486204
KBA05_ANTG1               1.445589
KBA05_ANTG2               1.012477
KBA05_GBZ                 3.341329
BALLRAUM                  5.114782
EWDICHTE                  3.981913
INNENSTADT                5.196747
GEBAEUDETYP_RASTER        4.060258
KKK                       2.776842
MOBI_REGIO                3.455386
ONLINE_AFFINITAET         2.390382
REGIOTYP                  4.700004
PLZ8_ANTG1                2.218213
PLZ8_ANTG2                2.952106
PLZ8_ANTG3                1.834660
PLZ8_ANTG4                0.865878
dtype: float64

Also, for the Underrepresented lot, the 4th, 7th and 3rd clusters are what should be considered¶

In [2979]:
# What kinds of people are part of a cluster that is underrepresented in the
# customer data compared to the general population?

cluster_4 = model.cluster_centers_[4]
pca_4 = pca.inverse_transform(cluster_4)
pca_4_2d = np.array(pca_4).reshape(1, -1)  # Reshape to 2D
underrepresented = scaler.inverse_transform(pca_4_2d)

overrepresented = pd.Series(data=underrepresented.flatten(), index=list(df.keys()))
display(overrepresented.head(45))

ALTERSKATEGORIE_GROB      1.858449
ANREDE_KZ                 1.813097
FINANZ_MINIMALIST         3.428030
FINANZ_SPARER             2.995665
FINANZ_VORSORGER          3.317424
FINANZ_ANLEGER            3.797548
FINANZ_UNAUFFAELLIGER     3.782066
FINANZ_HAUSBAUER          2.576793
HEALTH_TYP                2.040646
PRAEGENDE_JUGENDJAHRE     9.487140
RETOURTYP_BK_S            2.984537
SEMIO_SOZ                 2.588261
SEMIO_FAM                 3.902026
SEMIO_REL                 5.146993
SEMIO_MAT                 5.045156
SEMIO_VERT                2.614274
SEMIO_LUST                4.432167
SEMIO_ERL                 4.083262
SEMIO_KULT                3.721611
SEMIO_RAT                 5.147265
SEMIO_KRIT                5.566548
SEMIO_DOM                 5.657643
SEMIO_KAEM                5.348109
SEMIO_PFLICHT             4.998034
SEMIO_TRADV               4.802555
VERS_TYP                  1.496673
HH_EINKOMMEN_SCORE        2.735408
W_KEIT_KIND_HH            3.275667
KONSUMNAEHE               3.702680
CAMEO_INTL_2015          24.496840
KBA05_ANTG1               2.608642
KBA05_ANTG2               1.171091
KBA05_GBZ                 3.994225
BALLRAUM                  5.074914
EWDICHTE                  3.054661
INNENSTADT                5.599150
GEBAEUDETYP_RASTER        4.095579
KKK                       2.506862
MOBI_REGIO                4.101765
ONLINE_AFFINITAET         3.515877
REGIOTYP                  4.029731
PLZ8_ANTG1                2.914575
PLZ8_ANTG2                2.458729
PLZ8_ANTG3                1.020368
PLZ8_ANTG4                0.294508
dtype: float64

Discussion 3.3: Compare Customer Data to Demographics Data¶

  • From the diagrams we can conclude that customers are not universal to the general demographic. There is rarely any relationship between all even

The clusters where the mail order company should target are basically the 0nd, 8th and 9th clusters - overrepresented data¶

Using the 0th Cluster, these few set are denoted
  • People between ages 46-60
  • High money savers
  • Men
  • People who have low financial investments interest

The underrepresented data, as denoted in the 4th, 7th and 3rd clusters, tell a different story¶

Usint the 4th cluster specifically, being our most underrepresented few:

  • These are people between ages 30-45
  • The Female Gender
  • People who have average financial interests

Congratulations on making it this far in the project! Before you finish, make sure to check through the entire notebook from top to bottom to make sure that your analysis follows a logical flow and all of your findings are documented in Discussion cells. Once you've checked over all of your work, you should export the notebook as an HTML document to submit for evaluation. You can do this from the menu, navigating to File -> Download as -> HTML (.html). You will submit both that document and this notebook for your project submission.

In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]:
In [ ]: